Professional Documents
Culture Documents
A Thesis
Presented to
The Academic Faculty
by
Imon Chakraborty
In Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy in the
Daniel Guggenheim School of Aerospace Engineering
Approved by:
iii
ACKNOWLEDGEMENTS
It would seem that yet another graduate student’s Ph.D. journey has come to a
completion through several challenging, laborious, and/or time-consuming activities:
A research proposal was formulated, pitched, and approved, a research plan was
was successfully mounted. At the end of it all, I gladly and gratefully acknowledge
First and foremost, I would like to express my profound appreciation for the
members of my defense committee. I must thank whole-heartedly my advisor
Professor Dimitri Mavris for multiple things, among them the opportunity to be a
part of the Aerospace Systems Design Laboratory (ASDL) and the freedom to explore
admirably imperturbable patience, and shared many an enlightening insight into the
inner workings of academia and industry. I must also thank Professor Brian German
for finding time from his schedule on numerous occasions to review, critique, and
difficult for me to express in words how grateful I am to Dr. Elena Garcia for her
sustained patience, her ever-present willingness to review and critique my work, and
her constant reminders that I should remain mindful of the larger picture when the
engineer in me tended to dive at and strafe a pixel. Last but not least, I would like
to thank Professor Daniel Schrage and Dr. Ruben Del Rosario, who despite joining
the committee quite far along in the process, took time from their busy schedules to
iv
The graduate school experience is not merely about one’s research but also
about one’s interactions with colleagues, friends, and compatriots. I cannot possibly
name everyone, but let me at least name a few in no particular order: Kalyana
Gottiparthi, Daniel Garmendia and Burak Bagdatli (each of whom I’ve known since
and defense!), David Trawick (a fair bit of cross-country piloting together!), Michael
Miller, Metin Ozcan, Charlie Potter, Mohammed Hassan, and of course Gokcin Cinar
and Fatma Karagoz (continuous source of amusement!). To the few I mentioned and
to many others as well, I have this to say: my graduate school experience was made all
the richer through my interactions with all of you. I shall never forget the camaraderie
and I wish all of you the very best in your professional careers and personal lives.
Last but most certainly not least, I would like to thank my parents for the
and technically sound engineer that I have ever met, inspired me from a young age
power to ensure that I am steadfast in that pursuit. Collectively, they taught me that
there is no substitute for perseverance and no easy route to success, which in turn
cannot endure unless established upon a solid foundation of honesty and integrity.
Imon Chakraborty
Atlanta, GA
November, 2015
v
TABLE OF CONTENTS
DEDICATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . iv
SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxii
vi
III TECHNICAL APPROACH . . . . . . . . . . . . . . . . . . . . . . . 54
3.1 System, Subsystems, and Subsystem Architectures . . . . . . . . . . 54
3.1.1 Subsystem Architectures . . . . . . . . . . . . . . . . . . . . 55
3.1.2 Degree of Subsystem Electrification (DSE) . . . . . . . . . . 59
3.1.3 Total Fuel Impact and Total Weight Impact of Subsystems . 60
3.2 Integrated Sizing and Analysis Approach . . . . . . . . . . . . . . . 61
3.2.1 Definition of Design Requirements . . . . . . . . . . . . . . . 63
3.2.2 Traditional Aircraft and Engine Sizing Process . . . . . . . . 64
3.2.3 Generation of Subsystem Architecture Combinations . . . . . 67
3.2.4 Subsystem Architecture Sizing and Evaluation . . . . . . . . 76
3.2.5 Evaluation and Decomposition of Subsystem Impacts . . . . 79
3.2.6 Re-sizing of Aircraft and Subsystems . . . . . . . . . . . . . 86
3.2.7 Post-processing Analyses . . . . . . . . . . . . . . . . . . . . 90
3.3 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
vii
4.2.3 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.3 Nose-wheel Steering System (NWSS) . . . . . . . . . . . . . . . . . 117
4.3.1 Determination of Sizing Steering Moment . . . . . . . . . . . 117
4.3.2 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . 120
4.3.3 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.4 Wheel Braking System (WBS) . . . . . . . . . . . . . . . . . . . . . 121
4.4.1 Physical Modeling and Relationships . . . . . . . . . . . . . 122
4.4.2 System Sizing . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.4.3 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.4.4 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . 129
4.5 Thrust Reverser Actuation System (TRAS) . . . . . . . . . . . . . . 129
4.5.1 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . 130
4.5.2 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 131
4.6 Electric Taxiing System (ETS) . . . . . . . . . . . . . . . . . . . . . 131
4.6.1 Estimation of System Power Requirement . . . . . . . . . . . 132
4.6.2 Estimation of System Mass . . . . . . . . . . . . . . . . . . . 135
4.6.3 Estimation of Fuel Burn during Taxiing . . . . . . . . . . . . 136
4.7 Environmental Control System (ECS) . . . . . . . . . . . . . . . . . 138
4.7.1 Cabin Temperature, Pressure, and Airflow Requirements . . 138
4.7.2 Cabin Thermal Loads Analysis . . . . . . . . . . . . . . . . . 140
4.7.3 ECS Pack Model . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.7.4 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . 146
4.7.5 Drag Generation . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.7.6 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.8 Wing Ice Protection Systems (WIPS) . . . . . . . . . . . . . . . . . 149
4.8.1 Determination of Protected Surface Area . . . . . . . . . . . 150
4.8.2 Modeling Assumptions and Technology Assumptions . . . . . 155
4.8.3 Estimation of water impingement . . . . . . . . . . . . . . . 157
4.8.4 Estimation of Required Heat Flux . . . . . . . . . . . . . . . 158
viii
4.8.5 Determination of WIPS Sizing Flight Condition . . . . . . . 159
4.8.6 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . 161
4.8.7 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.8.8 Drag Penalty Estimation . . . . . . . . . . . . . . . . . . . . 164
4.9 Cowl Ice Protection System (CIPS) . . . . . . . . . . . . . . . . . . 164
4.9.1 Determination of Protected Surface Area . . . . . . . . . . . 165
4.9.2 Estimation of Required Heat Flux . . . . . . . . . . . . . . . 166
4.9.3 Determination of CIPS Sizing Flight Condition . . . . . . . . 166
4.9.4 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . 166
4.9.5 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.9.6 Drag Penalty Estimation . . . . . . . . . . . . . . . . . . . . 167
4.10 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
ix
5.3.5 Power Dissipation . . . . . . . . . . . . . . . . . . . . . . . . 201
5.4 Pneumatic Power Generation and Distribution System (PPGDS) . . 202
5.4.1 Precooler Heat Exchanger (PHX) . . . . . . . . . . . . . . . 203
5.4.2 Pneumatic Ducting . . . . . . . . . . . . . . . . . . . . . . . 204
5.4.3 Power Dissipation . . . . . . . . . . . . . . . . . . . . . . . . 206
5.5 Mechanical Power Generation and Distribution System (MPGDS) . 207
5.5.1 Mass Estimation . . . . . . . . . . . . . . . . . . . . . . . . . 207
5.5.2 Power Dissipation . . . . . . . . . . . . . . . . . . . . . . . . 207
5.6 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
x
7.3.3 Analysis of TFI Break-down for Different WIPS Solutions . . 247
7.3.4 Feasibility Analysis for Electric Taxiing System (ETS) . . . . 249
7.4 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
VITA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
xi
LIST OF TABLES
xii
28 Decomposition of All Electric subsystem architecture impact for SSA 243
29 Summary of model sensitivity parameters . . . . . . . . . . . . . . . . 258
30 Summary of technological SOTA K-factors . . . . . . . . . . . . . . . 267
31 Architecture impacts within Clusters 1 and 2 for SSA . . . . . . . . . 277
32 Architecture impacts within Clusters 1 and 2 for LTA . . . . . . . . . 278
33 Architecture impacts within Clusters 1 and 2 for VLA . . . . . . . . . 279
34 Best-performing SSA, LTA, and VLA subsystem architectures . . . . 280
35 Conventional, Reduced Bleed, and Zero Bleed engine off-take limits . 283
36 Impact of All Electric subsystem architecture for LTA . . . . . . . . . 303
37 Impact of All Electric subsystem architecture for VLA . . . . . . . . 303
xiii
LIST OF FIGURES
xiv
23 Geometry of brake rotors and stators . . . . . . . . . . . . . . . . . . 123
24 ETS mass addition as function of Maximum Takeoff Mass (MTOM) . 136
25 Cabin pressurization schedule . . . . . . . . . . . . . . . . . . . . . . 139
26 Component layout within thermodynamic model of ECS pack . . . . 143
27 Geometric parameters for definition of spanwise and chordwise EOP . 151
28 Wing leading-edge Extent of Protection (EOP) for some commercial
aircraft. All spans equalized to illustrate relative spanwise EOP. . . . 152
29 Spanwise EOP as a function of wing Mean Aerodynamic Chord (MAC)
for a number of commercial aircraft . . . . . . . . . . . . . . . . . . . 153
30 Pneumatic and electrothermal wing ice protection systems (WIPS). . 161
31 Simplified nacelle inlet geometry . . . . . . . . . . . . . . . . . . . . . 165
32 Electric system template - the basis for EPGDS architecture . . . . . 177
33 Schematic of conventional bleed system for turbofan engine . . . . . . 180
34 Schematic of conventional pneumatic system . . . . . . . . . . . . . . 182
35 Translation of logical subsystem connections to physical/geometric
equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
36 Conventional (pneumatic) ECS architecture (SSA-00000) . . . . . . . 213
37 Electric ECS architecture (SSA-00010) . . . . . . . . . . . . . . . . . 213
38 EPGDS architecture for electric ECS (SSA-00010) . . . . . . . . . . . 214
39 Summary of ECS electrification effects for SSA (SSA-00000 vs.
SSA-00010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
40 Comparison of thrust, TSFC, fuel flow, fuel burn, and vehicle weight
over the course of the mission (SSA-00010 relative to SSA-00000) . . 218
41 ECS mass flow and power requirements over mission . . . . . . . . . 219
42 Summary of ECS electrification effects for LTA (LTA-00000 vs.
LTA-00010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
43 Summary of ECS electrification effects for VLA (VLA-00000 vs.
VLA-00010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
44 Summary of effects of conventional to All Electric subsystem
architecture transition for SSA (SSA-73310 relative to SSA-00000) . . 222
45 Subsystem architecture snapshot generated by heuristic algorithm for
Airbus A340 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
xv
46 Subsystem architecture snapshot generated by heuristic algorithm for
Boeing 777 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
47 Subsystem architecture snapshot generated by heuristic algorithm for
Airbus A380 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
48 Subsystem architecture snapshot generated by heuristic algorithm for
Boeing 787 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
49 Inputs and outputs for truth model function call (executed as a
function call to NPSS with a MATLAB wrapper) . . . . . . . . . . . 237
50 Decomposition of total fuel impact for pneumatic and electric ECS into
contributions from mass, secondary power requirement, and direct drag
penalty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
51 Decomposition of total fuel impact for different WIPS solutions into
contributions from mass, secondary power requirement, and direct drag
penalty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
52 Electric taxiing feasibility analysis for SSA (SSA-00001) . . . . . . . . 250
53 Feasibility assessment of electric taxiing for LTA and VLA . . . . . . 251
54 Gross sensitivity of SSA-00000, SSA-70000, SSA-03310, and SSA-73310
to epistemic uncertainty K-factors . . . . . . . . . . . . . . . . . . . . 260
55 Impact of individual epistemic uncertainty K-factors on performance
of SSA-00000, SSA-70000, SSA-03310, and SSA-73310 . . . . . . . . . 261
56 Gross sensitivity of SSA-70000, SSA-03310, SSA-73310 and SSA-73311
to technological SOTA . . . . . . . . . . . . . . . . . . . . . . . . . . 268
57 Impact of individual technological SOTA K-factors on performance of
SSA-70000, SSA-03310, SSA-73310, and SSA-73311 . . . . . . . . . . 270
58 Comparison of 512 subsystem architectures for SSA, LTA, and VLA . 274
59 Definition of Candidate Subsystem Architecture Descriptor (repeated) 280
60 Performance comparison of SSA subsystem architectures using
scheduled penalty coefficients and architecture-to-engine association
logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
61 Mission performance comparison of selected SSA architectures from
Clusters 2 and 3 (RB engine) and Cluster 4 (ZB engine) . . . . . . . 286
62 Effect of electrification of actuation functions for SSA, LTA, and VLA
under the technology assumption that electric actuators for FCAS,
LGAS, NWSS, and TRAS weigh exactly the same as the hydraulic
actuators they replace . . . . . . . . . . . . . . . . . . . . . . . . . . 289
xvi
63 Summary of effects of conventional to All Electric subsystem
architecture transition for LTA (LTA-73310 relative to LTA-00000) . . 304
64 Summary of effects of conventional to All Electric subsystem
architecture transition for VLA (VLA-73310 relative to VLA-00000) . 304
65 Gross sensitivity of LTA-00000, LTA-70000, LTA-03310, and
LTA-73310 to epistemic uncertainty K-factors . . . . . . . . . . . . . 305
66 Gross sensitivity of VLA-00000, VLA-70000, VLA-03310, and
VLA-73310 to epistemic uncertainty K-factors . . . . . . . . . . . . . 305
67 Impact of individual epistemic uncertainty K-factors on performance
of LTA-00000, LTA-70000, LTA-03310, and LTA-73310 . . . . . . . . 306
68 Impact of individual epistemic uncertainty K-factors on performance
of VLA-00000, VLA-70000, VLA-03310, and VLA-73310 . . . . . . . 307
69 Gross sensitivity of LTA-70000, LTA-03310, LTA-73310 and LTA-73311
to technological SOTA . . . . . . . . . . . . . . . . . . . . . . . . . . 308
70 Gross sensitivity of VLA-70000, VLA-03310, VLA-73310 and
VLA-73311 to technological SOTA . . . . . . . . . . . . . . . . . . . 308
71 Impact of individual technological SOTA K-factors on performance of
LTA-70000, LTA-03310, LTA-73310, and LTA-73311 . . . . . . . . . . 309
72 Impact of individual technological SOTA K-factors on performance of
VLA-70000, VLA-03310, VLA-73310, and VLA-73311 . . . . . . . . . 310
xvii
List of Acronyms
AC Alternating Current
BF Block Fuel
CG Center of Gravity
DC Direct Current
xviii
EDS Environmental Design Space
HT Horizontal Tail
LE Leading Edge
xix
MAC Mean Aerodynamic Chord
MO Mixed Off-take
RB Reduced Bleed
xx
RHS Right Hand Side
SOTA State-Of-The-Art
TE Trailing-Edge
VT Vertical Tail
ZB Zero Bleed
xxi
SUMMARY
Traditional aircraft conceptual design focuses primarily on the sizing of the vehicle
and its propulsion system. The effects of the vehicle subsystems are largely accounted
for implicitly based on historical data and trends, and their detailed design is typically
relegated to subsequent design phases. Thus far, this has been possible largely due
solutions and architectures are considered, such as those proposed for future More
Electric Aircraft or All Electric Aircraft. Historical data regarding such subsystems
and their effect on the aircraft as a whole is either limited or non-existent. Moreover,
much more significant for such novel subsystems. In such cases, the incorporation
of a more thorough and explicit consideration of the aircraft subsystems into the
and analysis methods that are suitable for the early design phases with the traditional
aircraft sizing methodology. Suitable methods for early design phases are those
that are computationally inexpensive and do not require detailed aircraft definition.
The goal is to determine subsystem characteristics in sufficient detail to permit the
of subsystem architectures and any variations of these with aircraft size. This
xxii
objective is motivated by the observation that the progressive electrification of
subsystem architectures within the aerospace industry is occurring through More
Electric Aircraft, in which only some of the subsystems are electrified while others
The third and final objective is to assess the sensitivity of the predicted
from modeling limitations, modeling assumptions, and tool fidelity on one hand and
parameters to allow simultaneous sizing of the aircraft and its subsystems. The
subsystems are sized and evaluated using models that are computationally inexpensive
and require only limited input information that is either available or easily estimable
during early design. The connectivities among the different subsystem architecture
mission performance are used to simultaneously re-size the aircraft and its subsystems.
environment are evaluated by (i) assessing the aircraft-level and mission-level impacts
xxiii
off-take sensitivities generated using a higher-fidelity propulsion tool without direct
integration of the tool, and (iv) verifying that a subsystem architecture’s overall
secondary power requirements, and drag. Subsequently, the impact of epistemic and
subsystem sizing and analysis with traditional aircraft sizing, with a focus on
that provided the highest payoffs for each vehicle size. Further, it facilitated the
for the design of More Electric Aircraft which have already appeared in service and
xxiv
CHAPTER I
This chapter provides a brief overview of aircraft subsystem functions, the use
of different forms of non-propulsive power for subsystem operation, historical and
modern trends for subsystem architectures, and the effect of subsystems on aircraft
Aircraft subsystems, also known as Aircraft Equipment Systems (AES) [1], may
controllability, and comfort [2]. These include the Environmental Control System
(ECS), the Ice Protection Systems (IPS), the actuation systems for flight controls,
landing gear, thrust reversers, brakes, and nose-wheel steering, in addition to the
functionalities to the aircraft during the course of its flight or mission. For example,
the ECS is responsible for maintaining the cabin at levels of pressure, temperature,
and humidity that ensure sufficient human comfort despite possibly severe ambient
conditions. The IPS is designed to prevent or eliminate ice formation from external
surfaces of the aircraft that are at risk of suffering in-flight icing (such as leading edges
of the wings and engine nacelles). The flight control surface actuation subsystem
is responsible for ensuring the correct and precise movement or deployment of the
1
the maneuverability and flight characteristics necessary for a given condition or
mode of flight. This subsystem comprises primary control surfaces that are active
throughout the duration of the flight and secondary control surfaces which are used
the landing gear actuation subsystem must ensure the positive deployment, retraction,
and locking of the landing gear, while the thrust reverser actuation, brake actuation,
and nose-wheel steering subsystems are responsible for ensuring adequate deceleration
and control of the aircraft on the ground.
The primary function of the aircraft engines is to generate the thrust (power) required
to propel the vehicle in flight. This power is therefore often referred to as primary
main engines1 is to provide the power required by the subsystems to carry out their
termed secondary or non-propulsive power. When obtained from a gas turbine engine,
secondary power may be of the following forms:
the gas turbine engine. The pressure and temperature of the air thus obtained
is dependent on the compressor stage used and the engine throttle setting.
acceptable levels, the bleed air may be used for subsystems such as the ECS,
the IPS, hydraulic reservoir pressurization, etc.
1
On the ground, secondary power may also be provided by an Auxiliary Power Unit (APU). In
certain cases, the APU design may also permit it to supply secondary power in flight. Depending on
the magnitude of secondary power required, energy storage devices such as batteries (and perhaps
in future super-capacitors and fuel cells) may also be used.
2
2. Mechanical power: Also known as shaft-power off-take, mechanical power
may be extracted from the engine by driving a gearbox off the engine shaft.
Certain loads such as engine fuel and oil pumps may be driven by such an
converted to two additional forms of power for use by the aircraft subsystems:
(a) Hydraulic power: This is obtained by driving a hydraulic pump off the
psi and 4,000 psi are common for commercial and military applications
respectively, with newer commercial aircraft (e.g., Boeing 787) and military
aircraft (e.g., V-22 Osprey) making use of 5,000 psi systems [3].
generators off the engine gearbox, whose speed varies with engine
Drive (CSD), which can produce a constant speed output shaft from
DC loads are found in aircraft today, and both types of power (and the
pneumatic, mechanical, hydraulic, and electric. Since secondary power (in any form)
is extracted from the engine, there is an increase in the rate of fuel consumption of
the engine or, for the same engine thrust output, an increase in the thrust-specific
3
1.3 Gravitation Towards Conventional Subsystem
Architectures
Secondary power requirements for aircraft grew steadily from the Second World War
onwards, but even at that time, there was considerable debate regarding the optimal
means of providing secondary power for aircraft subsystems [5]. Some wartime aircraft
such as the British Short Stirling and the German Focke-Wulf Fw 190 made use of
electrical power for onboard functions [3], even though these aircraft did not have
powered flying controls. The British V-bombers of the 1950s (the Avro Vulcan, the
Vickers Valiant, and the Handley Page Victor), which had powered flying controls,
used electric power for actuation functions. The Vickers VC-10 of the 1960s also
used electric power for most flight controls. However, during this period, the power
densities of electric drive technologies did not develop at the same pace as those
of hydraulic systems, which were soon able to produce nominal system pressures of
3,000 psi. As this provided a simpler and lighter solution to the actuation problem at
the time, hydraulics became the mainstay for aircraft actuation functions for several
decades, and continue to have a significant presence on modern aircraft even today.
At the same time, pressurization and air-conditioning functions (ECS) and ice
protection functions (IPS) began to gravitate towards the use of pneumatic systems.
While the bleed air mass flow requirements increased with increasing aircraft size,
this was not considered a major problem for the turbojet engines or low bypass ratio
turbofan engines powering aircraft at the time, since sufficient core flow was available
Thus, the pattern of secondary power usage, i.e. the association of the aircraft
power for ECS and IPS functions, and (iii) hydraulic power for actuation functions [6],
4
Figure 1: The transition from conventional subsystem architecture to electric
subsystem architecture
have persisted for several decades. While the gravitation towards the conventional
subsystems [7], the persistence of such architectures has also been influenced by
the use of higher hydraulic system pressures (5,000 psi instead of 3,000 psi), and some
and pneumatic power (and relatively smaller amounts of electric power) have matured
over decades of aeronautical experience, but have now reached a so-called “technology
power electronics and electric drives [3, 8], has led to a renewed interest in electric
5
subsystem solutions as part of the More Electric Initiative [9]. The end goal is to
develop what has come to be called the All Electric Aircraft (AEA) [10, 11, 12], in
(Fig. 1, right). Since the technological risks involved with a direct single-step
preferred by the industry. This resulted in attempts to develop More Electric Aircraft
(MEA), intermediate steps in the possible transition to AEA that contained some
(but not all) electric subsystems. While it was acknowledged that this step-by-step
approach would initially obscure the true potential benefits of electric architectures,
it did agree with the aerospace industry’s preferred approach of risk mitigation and
and maintenance costs, turnaround times, range, and system reliability [4].
A major motivating factor for the MEI has been the recognition of the inefficiencies
example, represent significant energy wastage due to the need to reduce the
temperature and pressure of the extracted bleed air prior to supplying it to the
because despite the implementation of port selection logic, the bleed air temperature
and pressure are typically too high to be directly used downstream. The temperature
is typically above the auto-ignition temperature of fuel in fuel lines which may be
in close proximity to pneumatic ducts. Also, the high pressure presents challenges
for sealing and containment [13]. This gross mismatch between the available and
required (or usable) pneumatic energy [8] necessitates the use of a precooler and a
pressure regulating valve to reduce respectively the temperature and pressure, which
essentially amounts to simply rejecting extracted energy overboard. The progressive
trend towards higher bypass ratio engines also presents additional challenges as the
6
available core flow is reduced. With regard to this aspect, pneumatic IPS is a more
difficult problem to contend with, as the bleed requirements may equal or exceed those
for pneumatic ECS, but are not continuous (like for ECS) since the IPS can be turned
off at high altitudes [3]. As a result, electric power has been proposed for the ECS and
or “bleedless” architecture has already been achieved in the Boeing 787 Dreamliner,
in which electric power is used for the ECS and the wing IPS. Rather than using bleed
air from the engines, the Boeing 787 uses dedicated Cabin Air Compressors (CACs)
to pressurize the necessary quantity of external ram air to the required pressure for
supply to the cabin. Similarly, instead of using bleed air, the wing IPS uses electric
heating mats (or pads) bonded to the inner skin of the protected leading edge to
provide the necessary ice protection [14]. Since the ECS is the largest consumer of
non-propulsive power, a potential advantage arising out of the use of electric ECS is
the fact that the higher rating generators that are required can typically be motored to
perform the engine starting function [15]. This additionally eliminates the traditional
the disadvantage of being heavy (e.g., hydraulic pipework on an Airbus A300 sized
airplane weighs in excess of 800 kg [3]), use flammable and corrosive hydraulic
fluid, and are not necessarily immune to common cause failures despite triple
the landing approach phase, where high flow demands from the primary flight controls,
high-lift system, and landing gear must be supplied at low engine speeds [17]. At other
more benign flight conditions such as during cruise, when hydraulic flow requirements
are lower while engine speeds are higher, the pumps are under-utilized and operate
7
maintain pressurization in centralized hydraulic systems due to leakage flow. These
factors resulted in a renewed interest in the use of electric power for the actuation
functions, using electric actuators which would consume power only when the control
Electric actuation of flight control surfaces was given particular attention on the
military side [9]. The technology was especially attractive since the elimination of
a centralized hydraulic power source would reduce the vulnerability of the aircraft
in combat [18]. Actuation test programs were conducted on the C-141 Starlifter
and C-130 Hercules aircraft [19]. In the late nineties, NASA’s F-18 Systems
Research Aircraft (SRA) was used to test the performance of two types of electric
program [22] culminating in the F-35, the first production fighter to use electric
actuation for all flight control surfaces [23]. Commercial aviation has been more
cautious, but Airbus, after conducting technology maturation programs on the A320
and A330 aircraft, introduced two types of electric actuators in the Airbus A380 in
parallel to conventional hydraulics for several flight control surfaces [24]. For this
aircraft, a weight savings of 450 kg (1,000 lb) has been attributed to the shift from a
flight-critical controls such as spoilers [26, 27] and the high-lift system [28, 29].
Electric power for the remaining actuation functions has also been considered.
For example, the Airbus A380 features an Electric Thrust Reverser Actuation
System (ETRAS), while the Boeing 787 uses electromechanically actuated brakes [30].
Electric actuation for the landing gear extension/retraction and nose-wheel steering
functions was investigated as part of a research program [31], but is not found in any
8
A novel concept known as “Electric Green Taxiing System” (EGTS), where the
aircraft is taxied using APU-powered electric motors fitted to the main gear axles,
has also been tested successfully on the Airbus A320 aircraft through a collaborative
project between Airbus, Honeywell, and Safran [32]. This allows the aircraft to taxi
without the use of the main engines, and has been predicted to offer reasonable fuel
savings for aircraft operating mainly on short-haul routes. Other auxiliary benefits of
such a concept include the elimination of tugs and other ground personnel associated
with the push-back from the gate and a reduction in airport emissions.
Parallel to the proposed concept of the All Electric Aircraft is that of the All
Electric Engine (AEE) [6], where the only non-propulsive power off-take is in the
form of electrical power. The AEE is not a new concept either, and as early as
Provided a certain minimum rating was met, the generator could also be motored to
start the engine. Electric engine starting has been realized in the design of the Boeing
787, which uses Variable Frequency Starter-Generators (VFSGs) that can be motored
(powered) to start the engine (starting mode) and then driven by the running engine
to generate electricity (generator mode) [14]. Electrically powered engine accessories
(fuel and oil pumps) were also proposed, which would provide benefits in efficiency,
weight, size, and flexibility of speed control [34, 35]. An added benefit is a possible
reduction in required nacelle diameter as there is no longer the need for an accessory
gearbox, which has often caused an undesirable ballooning of the nacelle [15].
9
1.5 Subsystem Considerations during Conceptual Design
With some background regarding conventional and More Electric solutions for the
major aircraft subsystems having been presented in the preceding sections, the effect
of the subsystems on the aircraft sizing process is now addressed. While several
definitions of aircraft sizing exist in literature, the following is used in this work [36]:
so that the resultant aircraft should satisfy three criteria: matching power,
These two scales, the wing area (Sw ) and the required sea-level static engine thrust
(TSL ) may be considered along with the vehicle’s maximum takeoff weight (WT O ) as
being the three primary outputs from the aircraft sizing process.
The effect of the aircraft subsystems on the sizing of the aircraft is shown overlaid
on the refined sizing method of Raymer [37] in Fig. 2. As indicated in the figure, the
aircraft subsystems affect the conceptual phase sizing process in three ways:
empty weight of the vehicle through changes in the fixed equipment weight
3. Vehicle drag: The subsystems may contribute to the overall vehicle drag due to
the drag arising from ram air inlets or any necessary external modifications
aircraft fixed equipment weight, the conceptual phase designer has access to a vast
historical database of information that has been populated and updated over the
10
Figure 2: Effects of subsystems indicated on Raymer’s refined sizing method [37]
course of several decades of aeronautical experience. This database has also allowed
the regression of equations for the weights of various subsystems, which are quite often
of the form W = aX b . For example, the General Dynamics method [38] estimates
where q̄D is the dynamic pressure corresponding to the aircraft’s design dive speed.
Similarly, Torenbeek suggests the following relationship for the combined weight of
the hydraulic, pneumatic, and electrical systems2 based on aircraft empty weight We :
2
It is seen from Eq. 2 that the equation gives the combined weight of a number of subsystems,
and provides no means by which the individual masses of the subsystems can be determined.
Lammering [39] provides several additional examples of such inconsistent groupings in regressed
equations for conventional architectures.
11
The conceptual designer of AEA/MEA concepts, however, does not have access
to a similar historical database or regression equations for the weights of electric
among different subsystems had either been mild or non-existent. However, this may
not be the case for a fully or even predominantly electric subsystem architecture,
where a larger number of subsystems share the same form of energy: electricity.
The sizing cases for subsystem architecture components are also likely to be different
in this case. Efficiency improvements for conventional subsystems have generally
attempt to do the same for electric subsystem architectures would likely result in
a sub-optimal solution at the vehicle level. Thus, it would be fair to say that the
conceptual designer of AEA/MEA is faced with significant challenges, such as (i) the
Further, it is clear from the Boeing 787 and the Airbus A380 aircraft that
architectures will be a gradual one, likely featuring a range of More Electric Aircraft
be taken by the major aircraft manufacturers. This then leaves the MEA conceptual
designers with the additional task of determining which subsystems to target for
electrification. From the fact that the Boeing 787 and the Airbus A380 have different
electric subsystems, it is clear that there is no unique “right” choice. Rather, the
decision may very well vary with the aircraft category, and is of a multi-objective
nature that must be tackled in the face of considerable uncertainty and risk.
12
1.6 Previous AEA/MEA Feasibility Studies and Programs
Since the concept of AEA/MEA is not a new one, there is a reasonable amount
these, some notable independent efforts were undertaken in the Europe. A summary
technology suites considered involved that related to the AEA, i.e., the elimination of
power system. Weight and cost data were obtained using the Lockheed-developed
studying technology tradeoffs, the payload and missions were held fixed but the
airplane was re-sized in response to estimated equipment and fuel weight changes.
arising out of reduced systems weight, was avoided as the resulting designs would
end up with dissimilar payload/range capabilities. This tradeoff study considered the
incorporation of the various technology suites (including AEA technologies) into three
aircraft - a tri-jet high bypass turbofan 500-passenger subsonic airliner (based on the
L-1011-500 aircraft), and two short-haul twin-turboprop airplanes of 50-passenger
C, etc. into the baseline, comparisons were made as follows: baseline vs. baseline +
A, A vs. A + B, B vs. B + C, etc. All three aircraft considered were seen to benefit
13
from the AEA concept, but the effect on the 500-passenger aircraft was the most
substantial. The use of an all-electric secondary power system exhibited the most
impressive weight and cost payoffs. A major economic benefit was identified from the
elimination of bleed air for the ECS. The overall benefits comprised component and
in block fuel.
350-PAX aircraft (based on the 350-PAX, 5,000 NM Lockheed L-1011 aircraft), and
for each aircraft to evaluate mission fuel burn performance. The technologies were
grouped by functional similarity into eight main packages, two of which involved flight
control technologies and secondary power systems (both near-term and far-term).
Notably, the near-term secondary power system considered resulted in the elimination
of bleed air, while the far-term secondary power system also saw the removal of
centralized hydraulics. The study, which made use of the Lockheed-developed ASSET
tool, started with the establishment of the three baseline designs. Consultation
with in-house and vendor specialists was employed to obtain estimates of weights of
actuators, lengths and weights of wiring runs, etc. Weight deletions from the baseline
the baseline configuration. In this study, each tradeoff was performed by comparing
the tradeoff configuration to the original baseline that did not contain any of the
technologies. The claim was made that while in theory, benefits could not be added
together to ascertain combined benefits, in practice this could be done with negligible
error incurred. Like the 1980 Lockheed study preceding it, this study also noted the
fuel penalties incurred through the use of engine bleed air. However, it also recognized
14
efforts within the aerospace industry to pursue the reduction of bleed air requirements,
due to which the projected fuel savings of bleedless architectures would not be as
significant. While advanced secondary power technology showed large payoffs, it was
recognized that strong coordination would be needed in the development phase among
aircraft, engine, actuator, and avionics manufacturers. Like the preceding study, this
study also concluded that the payoffs (in terms of weight and cost savings) grew with
aircraft size.
In 1985, one of the two most prominent studies of the time was conducted by
NASA and Boeing [42]. Titled “Systems Study for an Integrated Digital/Electric
advanced electrical and digital systems for aircraft. The configuration selected for
use as a reference baseline was a modified Boeing 767 airplane powered by two
high bypass ratio E3 engines. This study noted the fact that the design of an
AEA would require the modification or selection of an engine for higher mechanical
power extraction. The engine considered in this study was meant to deliver high
shaft-power extraction (via the electrical generator) at low engine power with a low
extraction penalty compared to other existing turbofan engines. However, its bleed
air penalty was significantly higher than most other turbofan engines. Despite this,
the same engine was used to supply bleed air for the baseline aircraft customer
bleed requirements, even though it was clearly at an unfair disadvantage in doing
so [3]. Certain updated technologies were assumed for the baseline to represent a
1990 new airplane go-ahead. The baseline configuration was compared to the IDEA
assure system compatibility. For each subsystem, system options were evaluated
15
using trade studies, analysis, and equipment supplier consultation, to narrow down
the range of options to a finally-selected configuration. The IDEA configuration was
predicted to offer a 3 % reduction in fuel and a 1.8 % reduction in DOC assessed over a
1,000 NM segment relative to the baseline for the same payload/range conditions. The
system advances that were considered included all-electric secondary power extraction
(via generators only, with no bleed air and no direct shaft-power), electric engine
electrical power generation and engine starting, power distribution and management,
in addition to advanced data distribution and a digital flight control system. The
baseline for this study was a stretched-fuselage Lockheed L-1011 derivative carrying
350 passengers over 4,600 NM, with performance benefits evaluated at an average
stage length of 2,500 NM. Following the incorporation of electric technologies into
the aircraft and subsequent re-sizing, the study predicted large benefits that included
a 11.3 % reduction in block fuel and a 7.9 % reduction in direct operating cost. The
study utilized the Lockheed Advanced Systems Synthesis and Evaluation Technique
(ASSET). The IDEA concept featured the complete elimination of the hydraulic and
pneumatic systems. The E3 engine that was used for this analysis was based on a
bleed configured design, and was not cycle-optimized for shaft-power extraction. In
the absence of a specific cycle-optimized design, the authors stated that the physical
differences in the IDEA engines were computed parametrically from the baseline E3
design. It is noteworthy that while the summary of this study states that the block
fuel reduction was 11.3 % overall, only half of this was attributed to the incorporation
of all-electric secondary power systems. The remainder of the fuel burn benefit was
16
achieved through the assumption that an advanced computer-augmented fly-by-wire
flight control system could be incorporated that would allow relaxed static stability
Aircraft” was conducted by the NASA Lewis Research Center in 1985 [44]. This
study attempted to fill the gap left by two previous studies [40, 41] by analyzing
electromechanical ones for the flight control surfaces, landing gear, brakes, and
thrust reversers, and the removal of the centralized hydraulic system. Similarly,
the baseline pneumatic ECS was replaced with an electric motor-driven vapor cycle
system for temperature regulation and an engine gearbox-driven compressor for cabin
this study the existing specifications of the Boeing 767 aircraft were used with no
attempt made to adjust these figures for 1990s technology. The NASA/General
Electric Energy Efficient Engine (E3 ) was selected as the engine for the baseline
aircraft. The General Aviation Synthesis Program (GASP) was used for weight
sizing and mission performance analysis. This tool featured an integrated approach
in which changes in wing loading (for example) were propagated to connected design
parameters such as wing area, tail size, aerodynamics, etc. The study predicted an
uncycled weight saving of 2,950 kg (6,500 lb) through an empty weight reduction
reduction and the elimination of bleed air, a cycled weight saving in excess of 7,700
kg (17,000 lb), with a 9 % reduction in mission fuel burn. The most significant weight
reduction was reported to be from the electrification of the secondary power system
(i.e., removal of hydraulics and pneumatics), which allowed better load-sharing and
17
system utilization. The most significant fuel burn reduction was reported to be due
to the removal of engine bleed air for ECS functions.
Power Systems (CRISPS) study, which began in 1988, aimed to take a broader look
at future systems possibilities for a range of civil aircraft [3]. It involved a total of
baselines considered in this study was an Airbus A320 aircraft. The results of the
investigation are documented in an overall summary report [45] and an additional
short executive summary report [46]. In the latter, the estimated benefits were
presented along with their development risks (technical and financial). For the Airbus
A320 baseline, an AEA version was seen to offer approximately 4 % fuel burn savings,
which increased to almost 5 % when coupled with an AEE. It was predicted that
an AEE could result in significant benefits by eliminating the drag caused by the
accessory gearbox. However, both AEA and AEE were considered to be on the
as the IPS and actuation functions were considered to entail lower risks, but also lower
Development (2002-2006), and was tasked with exploring the way forward for
Goodrich Actuation Systems. The project’s goal was to identify, optimize, and
validate innovative aircraft equipment aimed at reducing non-propulsive power
consumption. The focus was on four main domains - engine systems, electrical power
18
systems, actuation systems, and pneumatic systems. The More Electrical Engine
featuring the embedding of electrical generators within the engine as the sole source
of secondary power (and removal of the accessory gearbox) was investigated. This
necessitated the electrification of the engine fuel/oil systems. The load analysis,
architecture, cabling mass, and stability for high voltage DC networks were studied.
To avoid the large losses associated with pneumatic ECS due to the down-regulation
requirement, a set of solutions ranging from the reduction of engine bleed air
pressure to the complete removal of the bleed system and a transition to electric
flight controls, thrust reversers, landing gear extension, and wheel braking was
evaluated. It was observed that the benefits arising out of removal of hydraulics
would be largest if multiple actuation functions such as flight controls and landing
gear were electrified [48]. Another significant aspect of the POA program was the
use of three interacting validation platforms: (1) an Engine Systems Validation Rig
(ESVR) for demonstrating the integration of electrical power generation and accessory
subsystems and the removal of the accessory gearbox, (2) Aircraft Systems Validation
Rig (ASVR), an electrical characterization test rig for validating electrical power
generation and distribution systems, actuation systems, and pneumatic systems, and
(3) Virtual Iron Bird (VIB), an integrated software environment for aircraft level
evaluation and optimization. This was linked with the ASVR in order to verify and
tune component models. Some of the conclusions from the program were that electric
technologies had potential for superior performance, but that when implemented
their potential benefit. The forum recognized that many new technologies were at
demonstration level and that many challenges and issues still needed to be addressed.
The increased weights of electrified systems were identified as a major challenge.
However, it was observed that these could be acceptable provided there was a net fuel
19
savings [48]. Finally, it recognized that functional thinking and cross-ATA expertise
was needed for true integration at the aircraft level [48, 47]. The POA program
and the architectural design space was rather limited. The detail of analysis that was
The More Open Electrical Technologies (MOET) project was a 36-month research
centralized hydraulics and all engine air off-takes. In line with the Advisory Council
for Aeronautics Research in Europe (ACARE) vision for 2020, MOET aimed at
establishing the new industrial standard for commercial aircraft electrical system
design. The target applications for technology infusion were the Airbus A320
short-range airliner, the Dassault Falcon 2000 business jet, the Alenia ATR-72
short-haul regional airliner, and also the Eurocopter Super Puma helicopter. The
project involved both the development of simulation models and also the development
and testing of hardware components using a variety of test rigs. The study concluded
when fully integrated at vendor and airframer facilities. Conceptual design studies
concluded that airplane level benefits in terms of maintenance, operational flexibility,
and technology growth potential were possible, without any fuel-burn performance
smaller short-range aircraft. The study found that the MEA short-range aircraft was
heavier than the reference (conventional) aircraft for the systems (control volumes)
considered. The MEA was also predicted to suffer a drag penalty due to the
incorporation of additional ram air inlets for the ECS. Due to the elimination of
20
the study predicted no net effect on fuel burn for the short-range MEA, likely due to a
mutual off-set between beneficial (reduced TSFC) and adverse (higher weight, drag)
effects. Certain specific topics were recommended for future projects, including only
one high voltage network (230 VAC or ± 270 VDC but not both), smart management
EMA-driven flight control surfaces and landing gear (specifically for regional jets).
One of the salient features of this effort was that a large volume of test data was
published and thereby made available in the public domain.
On the military side, the United States Air Force (USAF) led a program called
Integrated Vehicle and Energy Technology (INVENT ) whose main focus was to make
the aircraft and vehicle systems more energy efficient by maximizing overall system
program coined the term Energy Optimized Aircraft (EOA). Its efforts were aimed at
creation of a tip-to-tail thermal model for military aircraft, featuring Air Vehicle
System (AVS), Fuel Thermal Management System (FTMS), and Power Thermal
Management System (PTMS). Design space trades were explored by exercising the
PTMS model. The program noted that typical thermal, power, propulsion, and
vehicle systems had been designed and optimized at a subsystem level with little
resistance had made convective cooling difficult. One of the INVENT objectives
investigate aircraft system design space prior to validation work being attempted,
with the requirement that subsystem models (and their validity) be scalable over a
reasonable size/capability range, with the target being set at +/- 25 % with respect
to the design operating point. In particular, the ability of the model to simulate
21
upon. The difficulties arising out of proprietary tools and simulation environments
used by different subsystem manufacturers, which prevent aircraft-level integration
studies, was recognized [51]. An approach to address this issue that was developed by
the Air Force Research Laboratory (AFRL) and others is Distributed Heterogeneous
Simulation (DHS). This allowed subsystem models to remain within their native
Several recent dissertations and archival papers related to aircraft subsystems have
addressed different aspects of the problem in differing levels of detail. Some of the
relevant works are briefly described below, along with a discussion of their advantages,
most approaches (including this dissertation) are then identified, followed by a brief
statement regarding how this dissertation differs from others preceding it.
considering the case of the flight control surface actuation subsystem for a small
single-aisle aircraft with fixed configuration and design mission. A comparison
baseline values and their architecture was not considered further in greater detail.
Comparisons were made with regard to secondary power extraction and subsystem
22
In this work, the aircraft size was assumed to be fixed, and the effect of changing
subsystem architectures was captured through an increase in the aircraft range. The
effect of subsystem architecture on the sizing of the aircraft and its integration with
requirements for defining the conceptual architecture of MEA. The thrust of this
requirements-driven sizing were addressed in less detail. Only a fixed vehicle test-case
was considered, and the work did not attempt to consider the effects of subsystem
architecture on the sizing of the aircraft. Prior to this, Armstrong had addressed the
an interface for installation definition, and a method to defining the operating space
for the architecture. Spatial orientation and installation of components was taken
into account in this work. Armstrong considered limiting the alternative design
instead of incompatibilities.
considerations with the early design phases through a doctoral dissertation and
additional archival publications [56, 57, 58]. The author proposed generalized “power
system modules” which shared common interfaces for receiving global and local
parameters and propagating parameters such as mass and drag. While several
23
subsystem models were integrated by the author into the Airbus Mission and
Performance Tool (AMPT), an internal Airbus tool, the dissertation [56] provides
focus and insight on sizing and simulation for (i) wing ice protection, (ii) commercial
cabin systems, (iii) pneumatic power system, and (iv) electric power system. Further
reference is made to the work of Liscouet-Hanke in these four areas in relevant sections
(i) the development of a Graphical User Interface (GUI) to enable the analyst to
interact with the analysis environment and (ii) a preliminary analysis of the effect
of rejected heat loads from the aircraft systems using a global thermal model of
the aircraft (developed separately for another Airbus project). Additionally, the
works. However, the effect of re-sizing of the aircraft in response to subsystem impacts
was not explored in [56], in which the geometry, structure, and propulsion parameters
were kept fixed and only the impact of the subsystems (or power systems) architecture
was done in conjunction with Airbus as part of an internal project, and some of
presumably using proprietary data. Since this data is of a restricted nature, the
models themselves are not documented fully in the open literature, and consequently
the modeling approach in certain cases naturally lacks transparency. Likely due to
the same reason, most of the results presented are in a normalized or percentage
the author’s observation that several other works existed where the modeling
24
approach either required a high level of detail and was therefore suitable only for
a specific architecture, or where the use of restricted models and data limited the
and the analysis of the spatial layout of the architecture components was based on
fuel consumption penalties due to power off-takes but do not take into account
the engine type, operating condition, or power extraction architecture. For his
work he used instead multi-dimensional polynomial curve fits which were created
the aircraft and ascertain the so-called “snowball” effects of subsystem architecture
was also assessed. Some noteworthy trends included the effect of the following on
inputs such as ECS recirculation fraction and IPS extent of protection, (ii) aircraft
design parameters such as number of passengers and cabin dimensions, and (iii)
top-level aircraft requirements such as design range and payload. However, the
comparisons and sensitivities that were assessed were limited to a relatively small
Seresinhe [59] presented an electrical load sizing methodology suitable for early
design stages of large commercial aircraft, in which a generic baseline electrical load
architecture could be modified and adapted for a MEA architecture. While the
25
electrical load analysis and sizing of the power generation system were considered
with due diligence, the methodology was heavily dependent on valid estimates for the
electrical loads imposed by MEA subsystems being available as input. In case of the
electric ECS, the power requirement was computed from a simple thermodynamic
relationship, while for other subsystems such as the IPS and flight control surface
actuation, the required power was interpolated or scaled from the results of other
prior studies. In a separate but related work by the same author in the context of a
trajectory optimization problem [60], the effect of shaft-power and bleed air extraction
on the fuel consumption of the engine was investigated. Penalty coefficients for each
were determined as a function of flight altitude and Mach number. This work also
focused more on estimating electrical power requirement and did not account for
factors such as increased ram drag or weight for an ECS layout or increased weight
components was cited in order to justify this assumption. This work therefore did
since recent MEA investigations [49] have specifically identified the weights of added
Based on the conclusions and opinions expressed in the most notable studies which
architecture will allow the true benefits and potential of AEA to be realized.
An incremental piece-by-piece approach will at best yield a fraction of the
26
net potential benefit [10, 47]. Several studies, therefore, have compared
completely conventional subsystem architectures to all-electric architectures,
Airbus A380 and Boeing 787) indicates that industry prefers a progressive
4. In general, the majority of plausible solutions for each subsystem are not
incompatible with the majority of solutions for the other subsystems, even
vehicle subsystems with multiple possible solutions for each results in a large
combinatorial space of feasible/compatible subsystem architectures, all of which
5. A thorough investigation of the effect of vehicle class (i.e., size and weight)
technological risks.
not unfairly or overly penalized through the assumptions made. For example,
27
if electric actuation is being compared to centralized hydraulic actuation, a
state-of-the-art nominal pressure of 5,000 psi (which is already seen in service)
as such. The results of such comparisons are strongly dependent on the assumed
current state-of-the-art and its projected future evolution [61], especially with
electric actuator which was not necessarily of the same configuration, while the
8. The need to optimize the engine cycle to maximize fuel economy following the
substitution of bleed air requirement with shaft-power requirement has been
acknowledged [41]. However, it has also been noted that not all studies have
extraction is considered from an engine designed for mixed power off-take (both
it must be borne in mind that due to the considerable expense and development
28
possible increase in the utility of the vehicle through an augmentation of its
payload-range capabilities. The second is a further down-sizing of the vehicle
through re-sizing of the design, while maintaining the same payload-range and
point performance capabilities. The latter approach may be more applicable for
Due to the very nature of the subsystems analysis problem itself, certain
commonalities exist with regard to the approach followed in prior work performed
• The sizing of the subsystems is driven by the flow-down of top-level vehicle and
downstream.
• Once sized, the impact of subsystems mass, power off-takes, and drag penalties
at the aircraft and mission levels are assessed by linking this information with
a suitable aircraft sizing and mission performance analysis tool (which varies
from organization to organization).
The above characteristics are also to be found in the approach described in this
dissertation. However, what distinguishes this dissertation from other work is the fact
that it addresses certain aspects of the problem that were either not addressed by
29
1. A parsimonious modeling approach is undertaken using parameters that are
either available or reasonably estimated in the early design phase. Large-scale
2. The combinatorial problem arising from multiple solutions for each subsystem
in the magnitudes of shaft-power and bleed air that they are designed to supply.
30
Table 1: Data summary for Small Single-aisle Aircraft (SSA), Large Twin-aisle
Aircraft (LTA), and Very Large Aircraft (VLA) baselines. In each case, tabulated
data corresponds to an aircraft with conventional subsystem architecture (hydraulic
actuation, pneumatic ECS and ice protection, and no Electric Taxiing System )
Aircraft Identification
Aircraft data SSA LTA VLA
Passenger capacity 170 396 852
Design range (NM) 3,000 7,800 8,200
Cruise Mach number 0.785 0.84 0.85
Max. ramp weight (lb) 175,130 746,610 1,270,000
Sea-level static thrust (lbf) 2 x 26,244 2 x 114,220 4 x 69,872
6. The net impact of each architecture is broken down into the individual impacts
of each constituent subsystem, which are further decomposed into the individual
contributions from subsystem mass, secondary power requirement, and drag.
This can guide the focus of subsystem design refinement efforts by facilitating
conventional baseline.
31
7. The sensitivity of the performance of subsystem architectures is assessed with
respect to (i) variations in parameters that represent modeling uncertainty and
the elicitation of more insight than possible in some other works where only
conceptual designer with additional challenges due to the lack of historical information
subsystem architectures using only the limited information that is available during
aircraft conceptual design. This will permit designers to seek out combinations of
of possible relationships between vehicle size and the most favorable subsystem
architectures for that size. Further, in addition to allowing the designer to assess
the impact of a novel architecture relative to a conventional one, it will provide
traceability that permits the overall impact to be decomposed and attributed to its
underlying causes.
32
CHAPTER II
AND HYPOTHESES
In Chapter 1, it was established that the design of subsystems for commercial aircraft
for the effect of subsystems through heuristic rules and statistical regression equations
based off historical data. However, equivalent information of this nature is not
available for the electric subsystem architectures towards which the aerospace industry
is currently progressing as part of the More Electric Initiative. This therefore creates
significant challenges for designing aircraft with such subsystem architectures.
based on the observations listed above and those discussed previously, is presented in
§2.1. The research questions that must be addressed in order to realize this objective
are identified and presented in §2.2. The research questions, associated hypotheses,
and experiments which collectively form the research approach undertaken in this
dissertation are discussed further in §2.3 - §2.4.
33
2.1 Research Objective
The first capability allows initial sizing of aircraft subsystems and subsystem
architectures in parallel with aircraft and engine sizing during aircraft conceptual
design. Suitable sizing methods are those which are computationally inexpensive and
do not require detailed aircraft definition, thus facilitating rapid tradeoffs using only
particular subsystem architecture, but also the decomposition of the overall impact
into the contributions stemming from individual subsystems. This in turn allows the
respect to a suitable metric and for a given aircraft size to be identified. In addition,
since the predicted performance is influenced by the fidelity of the modeling approach
34
2.2 Statement of Research Questions
The stated research objective may be realized by addressing three major research
vehicle and mission level using computationally inexpensive sizing and analysis
methods that require only limited information regarding the aircraft design?
with respect to a suitably defined performance metric or figure of merit and what
variation in their performance, if any, occurs with variation in the aircraft size?
this chapter contains discussions leading up to the research questions and hypotheses
and overviews of the experiments that are addressed in greater detail in subsequent
35
Figure 3: Summary of research questions, hypotheses, and experiments
(ECS) solution that uses electric power may be compared against a conventional
ECS solution that uses pneumatic power. In this case, the electric solution
and the pneumatic solution are two competing solutions for the same subsystem:
the ECS. The comparison may also be between two differing approaches to the
same type of solution, e.g., electrothermal ice protection versus electroimpulse ice
36
factor in these examples is that the focus is a single subsystem whose design is being
varied, while the design of remaining subsystems is not altered unless necessary.
actuation, ECS, and ice protection systems may be compared to a design with
hydraulic actuation, and pneumatic ECS/IPS. In this case, the comparison is between
architecture and a conventional subsystems architecture. This differs from the first
type since the designs of multiple subsystems are being varied simultaneously.
For both the first and second comparison types, multiple choices exist regarding
the metrics chosen for the comparison. A subsystem-level metric such as the net
mass or weight of the subsystem ultimately feeds into a vehicle-level metric such as the
aircraft’s Operating Empty Weight (OEW). The OEW directly affects a mission-level
metric such as required fuel, which in addition is also affected by subsystem secondary
power requirements and drag increments. The OEW and the fuel requirement in
turn affect other mission-level metrics such as Maximum Takeoff Weight (MTOW) or
Maximum Ramp Weight (MRW). The intent and scope of the comparison determines
methods is an important factor to consider, since the conceptual aircraft design phase
is dominated by the need to perform rapid trade studies or comparisons between
The success of the sizing and analysis methodology is also contingent on the input
parameters required for the subsystem sizing methods remaining limited to those
that are either available or easily estimated during conceptual aircraft design. For
example, a method whose evaluation requires knowledge of wing planform area and
sweep is acceptable, but one that requires detailed definition of the wing’s airfoil or its
37
pressure distribution is not, since these latter parameters are finalized only through
rigorous optimization occurring further downstream in the design process. These
vehicle and mission level using computationally inexpensive sizing and analysis
methods that require only limited information regarding the aircraft design?
(a) the initial sizes and characteristics of major components in individual power
(b) the initial sizes and characteristics of additional components in power generation
and distribution subsystems can be determined based on the identified requirements
(c) the mission performance analysis reflects the direct and indirect effects of the
38
(d) relevant dependencies can be established among subsystem-level and aircraft-level
parameters to allow the simultaneous re-sizing of both parameter sets based on
example of the functional requirements mentioned in statement (a), the ECS may
the determination of the necessary characteristics of ECS packs and their constituent
elements such as electric generators and hydraulic pumps, which serve as the sources
or producers of electric and hydraulic power respectively. The sizing conditions for
these components are typically encountered only briefly during a normal flight, or in
some cases they are encountered only during failure or degraded operating conditions.
This statement may be tested by verifying whether such conditions can be identified
and characterized using the information gained through (a) in addition to relevant
39
heuristic rules or guidelines available to the conceptual designer.
The direct and indirect effects of the subsystem architectures on the propulsion
system performance (statement (c)) originate from the incremental fuel flow rate due
ẇf = Treq · SF C
Here Treq and SF C are respectively the required thrust and the engine thrust-specific
fuel consumption for a given flight condition, ∆SF Cspx and ∆SF Cbx are the
extraction respectively, and ∆D0 = q̄Sw ∆CD0 and ∆Di = q̄Sw ∆CDi are respectively
the increments to zero-lift and induced drag components (or coefficients). To account
for the direct effects of subsystem secondary power extraction on the engine (either
∆SF Cspx = f (Pspx ), ∆SF Cbx = g(ṁbx ). The subsystems also indirectly affect the
required fuel flow rate by affecting the required thrust through the addition of drag to
the vehicle. The lift-independent drag component ∆D0 captures the effect of external
modifications to the aircraft through the incorporation of ram air inlets and the
associated momentum drag of admitted ram air. The lift-dependent drag component
∆Di accounts for the increase in induced drag that occurs due to the increased lift
required to offset the added mass and fuel requirement of the subsystems.
the second group and vice versa. For example, let it be assumed that the change
40
from conventional to electric ECS architecture results in a reasonable change in the
vehicle’s OEW. In response to this change, the wing planform area Sw may need to
be changed. This in turn may result in changes in the dimensions of the leading-edge
and trailing-edge control surfaces. These may then cause changes in the weights of
the wing IPS and control surface actuators. These latter two changes, triggered by
changes in the architecture of an unrelated subsystem, will then feed back into the
vehicle’s OEW, thus creating a circular dependency. This necessitates a set of aircraft
re-sizing rules for such scenarios, which will be discussed in a subsequent section. To
• Experiment 1.1 (§ 6.1): Demonstrate that the developed sizing and analysis
metrics of interest, while assuming fixed vehicle size (no re-sizing of the vehicle)
and invariant design for all unaffected subsystems.
two competing solutions are the conventional ECS and the electrified
ECS. Based on the solution chosen, the ECS mass, secondary power
41
architectures using relevant vehicle-level and mission-level metrics of interest.
Additionally, re-size the aircraft and demonstrate that the overall impact can
Protection System (WIPS), and Cowl Ice Protection System (CIPS). The
(ESA). For both cases, the aircraft are re-sized to meet the same point
The two most notable commercial MEA in service today, the Airbus A380 and
the Boeing 787, show important differences with respect to the subsystems that
some flight control surfaces. An Electric Thrust Reverser Actuation System (ETRAS)
is also present. However, other subsystems such as the ECS and the IPS retain a
42
conventional architecture. On the Boeing 787, the extent of electrification of the
subsystem architecture (in terms of secondary power consumption) is greater. The
most significant electrification is clearly that of the ECS, followed by the wing IPS.
are used. Additionally, a few of the wing spoilers and the Trimmable Horizontal
Stabilizer (THS) are electrically actuated. There is more significant usage of localized
hydraulic circuits that are pressurized by Electric Motor (Driven) Pumps (EMPs)
rather than Engine Driven Pumps (EDPs). The immediate question that arises based
on these observations is: Why do these two More Electric Aircraft differ in both the
A possible and quite likely explanation for this particular case may be
architecture shows enhanced performance and (ii) variations in the performance with
aircraft size. These therefore form the basis for the second major research question:
with respect to a suitably defined performance metric or figure of merit and what
variation in their performance, if any, occurs with variation in the aircraft size?
1. As there are multiple subsystems and multiple solutions for each, there is a large
43
elements present within them. Since these elements contribute directly to
subsystem mass, a means to determine which elements are necessary and how
they are connected to one another is required before the sizing of the elements
mixed secondary power in the form of both shaft-power and bleed air. If a
the ability to capture the effect that the subsystem architecture off-takes would
measured with respect to a suitably defined figure of merit) with aircraft size may
be ascertained through the integrated analysis if it can be demonstrated that:
requirements
44
(b) with an estimate of the type and magnitude of non-propulsive power requirements,
it is possible to capture the effect of changes in the engine cycle that would be
the top-level figure of merit can be identified and decomposed further into
contributions from subsystem mass, secondary power requirements, and drag
increments
whether certain architecture elements are required or not. For example, if actuation
functions use hydraulic power, then it is clear that hydraulic systems and pumps
would depend on how flight-critical the hydraulically actuated functions are. This
determination typically occurs later in the design process using Failure Modes and
Effects Analyses (FMEA) or Fault Tree Analyses (FTA). Statement (a) makes the
claim that an equivalent determination that is suitable for the purpose of early
analysis tool. It is also possible to integrate such a tool directly with the subsystems
sizing and analysis environment. However, this may be infeasible from the point of
view of computational expense. Instead, a higher-fidelity propulsion system analysis
tool can be used to establish a truth model, which can be queried in order to determine
the effect of secondary power off-takes on the engine’s performance. Experiment 2.2
45
Statement (c) is concerned with defining a suitable performance metric for each
subsystem to be sized to based on its effects on vehicle level energy consumption.
There is potentially a difference between this approach and designing each subsystem
fuel-consuming aircraft, the effect of the subsystems on the energy optimality of the
vehicle may be regarded as the increment in fuel consumption due to their presence
Thus, each subsystem will penalize (increase) the vehicle’s rate of fuel consumption
(fuel flow rate) through its shaft-power requirement (if present), bleed air requirement
(if present), direct drag increment (if present), and weight (always present). The
absolute and relative magnitudes of these components will, of course, vary from
of the additional power required to support (operate and carry) the subsystems. To
obtain the additional energy required to support the subsystems, the evolution of the
The relative magnitudes of the four effects (shaft-power, bleed, drag increment, and
46
characteristics of the subsystems. Considering for example the landing gear actuation
subsystem, which may use either hydraulic or electric power, Pspx 6= 0, ṁbx = 0.
actuator alone, ∆D0 ≈ 0. However, the landing gear actuator operates twice per
flight, for about 20-30 seconds per occurrence, whereas its weight is borne within the
aircraft for the entire flight. Considering the integration over time in Eq. 5 and then
Eq. 6, it is clear that κspx ≪ κw . In other words, based on the manner in which it
impacts a mission-level metric such as fuel consumption, there is strong justification
for concentrating design effort on minimizing the actuator weight. Similar arguments
apply to subsystems which operate infrequently (i.e., do not consume secondary power
continuously) and essentially serve as dead-weight for the majority of the flight.
ṁbx 6= 0 or Pspx 6= 0. In both cases Wsub 6= 0, D0 6= 0, the latter since the use of
ram air as the heat sink adds drag to the vehicle. Since the ECS operates throughout
the flight, it is much less clear a priori what the relative magnitudes of the factors
κspx , κbx , κd , and κw are, and whether any of them is dominant. Further, considering
a potential transition from conventional to electric ECS architecture, the bleed air
requirement ṁbx is replaced by shaft-power requirement Pspx , but the drag penalty
∆D0 is also higher as the cabin air must now be admitted through a separate ram
air inlet. It follows then that for an electric ECS, the magnitudes of the factors κspx ,
κbx , κd , and κw , and thus the focus of design efforts may be considerably different.
In the case of an optional subsystem, which may or may not be present within
an architecture, a determination of whether it can “buy its way onto the airplane”
must be made. For example, an Electric Taxiing System (ETS) permits the aircraft
to be taxied without using main engine thrust. The practicality of such an optional
subsystem may be evaluated through a modified form of Eq. 5,
✟✯0
✟ ✟✯0
✟
∆wf = ∆wf,spx + ✟
∆w f,b + ✟
∆w f,d + ∆wf,w − ∆wf,mc , (7)
47
in which ∆wf,mc > 0 represents any fuel saved due to mitigating circumstances. In
this case, this would be due to the fact that the electric motors are powered from
the APU while the main engines are not turned on. The desired outcome from this
optional subsystem is fuel savings: ∆wf < 0. It is clear that the sign of ∆wf is
(function of motor power draw), and ∆wf,w (function of additional weight). Further,
since electric taxi motors become dead-weight during non-taxiing phases of flight, it
is clear that the longer the duration of the flight ∆t = tf − ti , the less attractive this
feature will become. In fact, if for very long-haul flights, ∆wf > 0 is obtained, then
the inclusion of this technology may not be justifiable. Such conclusions were drawn
by an Airbus trade study into the Electric Green Taxiing System (EGTS) [32].
Experiments 2.1, 2.2, and 2.3 are intended to test the statements of Hypothesis 2,
while Experiment 4.0 is intended to address Research Question 2 directly. The formal
statements of all the experiments are as follows:
48
of redundancy. This is done by inspecting and comparing the association
of (i) power systems to power consumers, (ii) power systems to power
– Overview: A performance model for a gas turbine engine sized for a given
secondary power extraction using a higher-fidelity propulsion system sizing
and analysis tool is treated as the truth model. The truth model is queried
subsystem’s mass, secondary power requirements, and drag increments (if any)
subsequent design efforts for the subsystems. Further, verify whether this
presented. The net impact is first decomposed into contributions from the
49
from each subsystem’s mass, secondary power requirements, and drag
increments. (ii) The utility of the two-level decomposition in determining
The discussion thus far has focused on (i) sizing the aircraft subsystems using
the limited information that exists regarding the design in the conceptual design
phase, and (ii) assessing the impacts of subsystem architectures on aircraft-level and
mission-level metrics. However, the performance predicted for an architecture may
be quite sensitive to uncertainties from various sources. For the purposes of this
50
2. Technology uncertainty: This exists due to the inability to predict with 100 %
confidence the technological state-of-the-art (SOTA) at a future time
A few examples of epistemic uncertainty are provided as follows: (i) The actuation
loads for certain control surfaces are determined based on their dimensions, the
estimation of their hinge moment coefficients, and the identification of the flight
only a preliminary estimate (at best) can be made for each of these, and therefore
the computed actuation load is subject to epistemic uncertainty. (ii) The computed
power (heat) requirement for the IPS depends on an estimate of the surface area to be
protected and the net heat flux for a given set of atmospheric and flight conditions.
Without detailed icing simulations, only a preliminary estimate can be made for
the protected area, while the computation of the heat flux is subject to multiple
simplifying assumptions. (iii) The computed total ECS thermal load depends on an
estimate of the internal cabin heat loads and the rate of heat exchange between the
cabin and the ambient. Without detailed cabin thermal simulations, only preliminary
Technology uncertainty, on the other hand, arises as the rate of evolution of the
certainty. Such predictions are often required for feasibility studies or due to the fact
that the design process for a complex commercial aircraft may span over a decade,
thus necessitating a projection of the technological SOTA to correspond to the time
of fabrication. This is even more relevant for the case of More Electric subsystem
electric motors, power electronics, etc., which are on reasonably steep development
curves. The presence of these two forms of uncertainty motivated the third major
research question, which is stated as follows:
51
• Research Question 3: What is the sensitivity of the predicted performance of
subsystem architectures to (i) variations in model parameters chosen to reflect
electric motors, power electronics, and generators, which have a significant impact
include the following: (i) conventional, (ii) All Electric, (iii) bleedless (with
52
of certain key internal model parameters is determined. A determination
is also made regarding whether some architectures are more/less sensitive
include the following: (i) All Electric, (ii) bleedless (with conventional
actuation), and (iii) electric actuation (retaining pneumatics). The
This chapter introduced the research objective of this dissertation and identified the
experiments were outlined whose intended purpose is to test the hypothesis statements
or address the research question itself. The technical approach pursued in this
and 5. The experiments aimed at verifying the capabilities of the developed approach
and assessing the performance and sensitivities of subsystem architectures follow
of the research effort are discussed in Chapter 10 along with the identification of
53
CHAPTER III
TECHNICAL APPROACH
During the traditional aircraft conceptual design process, only limited attention has
been given to the aircraft subsystems, which instead are considered in more detail
in subsequent design phases [37]. However, from the discussion in previous chapters,
it is clear that this can no longer be the case for novel or unconventional subsystem
architectures that may be used for future AEA or MEA. In fact, for the design of such
Chapter 2, this was identified as one of the primary objectives of this dissertation.
This chapter contains technical details regarding the approach taken to meet the
research objectives. The relationship between system and subsystem within the
the subsystems considered within the scope of this work, a measure of subsystem
are first presented in §3.1. Following this, the developed integrated sizing and analysis
approach is presented, and each of the modules is discussed in some detail in §3.2.
system functionality. Since the terms system and subsystem are ubiquitous within this
dissertation, it is necessary therefore to define what constitutes the system and what
constitutes subsystems. For the purpose of this dissertation, the system is the aircraft
itself, with the Aircraft Equipment Systems (AES) constituting the subsystems.
54
Figure 4: Generalized representation of subsystems architecture
While the term subsystem suggests a certain amount of physical and functional
discussed within the remainder of this dissertation, are assumed to fall within the
of secondary power. By this definition, the power sources include (i) electric
generators, (ii) hydraulic pumps, and (iii) pneumatic bleed ports. In each case,
these are the points within the subsystem architecture where the corresponding
55
form of secondary power first becomes available. Multiple power sources of the
same type may be present (e.g., multiple pumps and generators)
3. Power systems: These are responsible for receiving secondary power, and
may occur, (ii) hydraulic systems, within which hydraulic pressure may be
regulated, and (iii) pneumatic systems, within which elements may be present
that regulate the state of the pneumatic power received. Like the power sources,
the subsystem architecture may feature multiple power systems of the same type
4. Power distribution elements: These link the power sources to the power
systems and the power systems to the power consumers. They are responsible
for physically conveying the appropriate secondary power from one point to
another. They include (i) electrical cabling/wiring, (ii) hydraulic piping, and
(iii) pneumatic ducting. Connecting elements may link multiple power sources
to the same power system, and multiple power systems to the same power
consumer
The subsystem architecture components when presented in the order above also
represent the actual flow of secondary power within the subsystem architecture.
is required. In other words, the power consumers must be addressed first, which
involves the determination of their secondary power requirements, mass, and direct
56
drag contributions, if any. These are determined based on a flow-down of requirements
from the aircraft level and their translation into corresponding subsystem-level
requirements.
Within this dissertation, the power sources, power systems, and power distribution
elements for a particular secondary power type are collectively referred to as the Power
Generation and Distribution Systems (PGDS). The sizing of the PGDS elements
follows that of the power consuming subsystems, and requires knowledge regarding not
only the magnitudes of secondary power required by the power consuming subsystems,
but also the connectivity of elements within the subsystem architecture. The
subsystems that are addressed within the scope of this dissertation are summarized
in Fig. 5. The power consuming subsystems considered include the (i) Flight
Controls Actuation System (FCAS), (ii) Landing Gear Actuation System (LGAS),
(iii) Nose-wheel Steering System (NWSS), (iv) Wheel Braking System (WBS),
(v) Thrust Reverser Actuation System (TRAS), (vi) Electric Taxiing System (ETS)1 ,
(vii) Environmental Control System (ECS), (viii) Wing Ice Protection System
A PGDS was considered for each of the secondary power types. Further,
since hydraulic and electric power are obtained from the conversion of mechanical
power (shaft-power), four PGDS were considered in total: (i) the Electric Power
1
Of the power consuming subsystems, the Electric Taxiing System (ETS) differs from the others
in that it is an optional subsystem that need not necessarily exist within all subsystem architectures.
57
Figure 5: Summary of power consuming subsystems and power generation and
distribution subsystems considered within the scope of the dissertation
Generation and Distribution System (EPGDS), (ii) the Hydraulic Power Generation
and Distribution System (HPGDS), (iii) the Pneumatic Power Generation and
Distribution System (PPGDS), and (iv) the Mechanical Power Generation and
Since the All Electric Engine (AEE) is not considered within the scope of this
dissertation, the MPGDS and the EPGDS exist within all subsystem architectures
Of the power consuming subsystems, the FCAS, LGAS, NWSS, WBS, and
TRAS are frequently referred to within this dissertation as the actuation functions
or the actuation subsystems, since they are involved in the actuation of movable
58
subsystems architecture, they would use electric power. These two possible secondary
power types for the actuation functions are indicated by the arrows in Fig. 5.
Similarly, the ECS, WIPS, and CIPS within a conventional architecture use pneumatic
power, while within an electrified architecture, they too would use electric power.
The fact that multiple secondary power options exist for these subsystems is the
basis for the existence of numerous combinatorial possibilities for the aircraft-level
subsystems architecture. Further, multiple solutions may exist for each subsystem
that make use of the same secondary power type, which results in a further
2. Merely considering the installed electrical power generation capacity or its ratio
to hydraulic and pneumatic capacities does not directly indicate the degree to
3. The metric must account for the fact that certain subsystems operate
throughout the flight while others are only active for brief durations
The DSE is defined as the ratio of the electrical energy consumed by N power
59
energy consumed by them over the same duration in either electrical, hydraulic, or
pneumatic form.
PN R T
i=1 0
Pelec (t).dt
DSE = PN R T (8)
i=1 0
{Pelec (t) + Phyd (t) + Ppneu (t)} .dt
Ppneu (t) = ṁb (t) h(Tref ), where h(Tref ) is the specific enthalpy of the bleed air at
a reference bleed air temperature, taken here to be Tref = 500 K (based on typical
bleed system temperatures given by Hunt [13]). From Eq. 8, it is clear that DSE = 1
indicates an AEA in which power consuming subsystems use only electric power.
power (installed pump capacity), and pneumatic power (maximum rate of bleed air
extraction), since it is well-known that the installed capacity is not completely utilized
A subsystem penalizes the vehicle’s fuel burn through weight, the extraction of
secondary power from the engine in the form of shaft-power and/or bleed, and the
direct generation of additional drag. The summation of these effects over a defined
The Total Weight Impact (TWI) may additionally be assessed by adding the
60
The TFI of a subsystem essentially indicates its contribution to the aircraft’s
mission fuel burn, while the TWI essentially indicates its contribution to the ramp
weight. Moir [63] states that when comparing systems, the overall system weight
penalty (i.e., the TWI) should be used. However, in case of More Electric subsystem
stated goals of the More Electric Initiative is to reduce fuel consumption, it would
unfavorable TWI. Therefore, while both TFI and TWI are computed for subsystem
A top-level overview of the implemented approach for the integrated sizing and
involves the flow of necessary information to/from the following major modules, each
7. Post-processing analyses
61
62
These are the standard design requirements that apply to the design of commercial
is shown in Fig. 7
63
integration with main lifting surfaces, the disposition of major components,
empennage and landing gear configuration, etc. This dissertation is limited to
The design requirements for the aircraft are assumed to be known. Further, it
is assumed that (i) the design requirements are invariant to the choice of subsystem
architecture for the aircraft, and (ii) there is no extra credit for a design whose
The mission and operational requirements and the notional air vehicle concept drive
the traditional aircraft and engine sizing process. The objective of this process is to
obtain a geometric scale for the vehicle in terms of the wing planform area Sw , a
propulsive power scale in terms of the required sea-level static thrust TSL , and the
1. Power (or thrust) matching: balancing the available power (or thrust) against
the required power (or thrust) to satisfy the point performance requirements
2. Energy (or fuel) matching: balancing the available energy (or fuel) against the
3. Volume balance: balancing available volume against required volume. For the
64
conventional aircraft configurations that are analyzed here, this is often satisfied
implicitly. Therefore, the focus is on the first two items
propulsion, and weight estimation. On the aerodynamic side, the notional concept
allows for an initial estimate of the lift and drag characteristics of the aircraft in
terms of the lift curves (lift coefficient CL versus angle of attack α) and drag polars
(drag coefficient CD versus lift coefficient CL ) for clean and flapped configurations.
The propulsion side starts with an estimate of the engine specific fuel consumption
(SF C), and variations of this quantity and the engine thrust with altitude and Mach
number. In addition, there may be engine scaling laws that allow for a preliminary
estimate of dry engine weight. The weight estimation side begins with an empirical
Since the weight of the aircraft WT O is not known at this point, the thrust
balance (ensuring that available thrust exceeds the thrust required to satisfy point
performance requirements) may be performed through constraint analysis. Here the
goal is to determine a thrust-to-weight ratio TSL /WT O and wing loading WT O /Sw
that is in the feasible region formed by the superposition of the point performance
(11)
W T
in which β = WT O
is the ratio of current vehicle weight to WT O , α = TSL
relates the
current minimum required thrust to the equivalent thrust at sea-level static (SLS)
conditions, q̄ is the dynamic pressure, K1 , K2 , and CD0 are the coefficients of the
parabolic drag polar CD = K1 CL2 + K2 CL + CD0 , and n is the load factor [64].
making suitable substitutions and simplifications to Eq. 11, each point performance
65
Figure 8: Securing power/thrust balance using constraint analysis to satisfy point
performance requirements
f (WT O /S). Considering all such constraint curves simultaneously allows the feasible
The fuel balance (ensuring that the available fuel is greater than or equal to the
required fuel to complete the mission requirements) may be secured through mission
performance analysis. The fuel requirement for the sizing mission may be expressed
as a mission fuel fraction Wf /WT O , which in turn may be expressed as the product
where the fuel fraction for segment ‘k’ is the ratio of aircraft final weight at the
(k) (k)
end of the segment (Wf ) to initial weight at the beginning of the segment (Wi ).
Raymer [37] provides expressions for preliminary estimates of the segment fuel
fractions that may be used to evaluate Eq. 12. With known weights for the payload
(Wpl ) and crew (Wcrew ), the takeoff weight WT O may be obtained through an iterative
convergence of
Wpl + Wcrew
WT O = , (13)
1 − We /WT O − Wf /WT O
66
where We /WT O is the empty weight fraction of the vehicle. While a first estimate
of this fraction may be made based on historical trends, a better estimate may be
Several aircraft sizing and performance analysis codes already exist that perform
the activities described above (in Chapter 1, several such tools were mentioned
analyses). This dissertation uses a tool developed by NASA Langley Research Center
called Flight Optimization System (FLOPS) [65] for this purpose. If the design
requirements are provided as inputs, then the tool is capable of determining the
vehicle size, gross weight, and fuel necessary to satisfy these requirements. The
weight buildup relationships within this tool (as with most other similar tools) are
For both conventional and More Electric subsystem architectures, there may be
multiple methods of achieving the required subsystem functions. For example, the
actuators used for the ailerons, elevators, rudders, and spoilers may be classified
according to the type of power that they consume: conventional actuators consuming
classified based on the nature of the output motion: linear actuators or rotary
the kinematics of the flap mechanism, may be linear or rotary), or using centralized
67
Table 2: Matrix of Alternatives (MoA) based on some subsystem solutions
Subsystems Options #
Ailerons Lin-hyd, Rot-hyd, Lin-EHA, Lin-EMA, Rot-EMA 5
Elevators Lin-hyd, Rot-hyd, Lin-EHA, Lin-EMA, Rot-EMA 5
Rudder Lin-hyd, Rot-hyd, Lin-EHA, Lin-EMA, Rot-EMA 5
Spoilers Lin-hyd, Rot-hyd, Lin-EHA, Lin-EMA, Rot-EMA 5
Flt. Ctrl. LE-flaps Lin-hyd, Rot-hyd, Lin-elec, Rot-elec,
H-PDU, E-PDU, H/E-PDU 7
TE-flaps Lin-hyd, Rot-hyd, Lin-elec, Rot-elec,
H-PDU, E-PDU, H/E-PDU 7
THS H-jackscrew, E-jackscrew, H/E-jackscrew 3
power drive units (PDUs), which in turn may be powered hydraulically, electrically,
or with one hydraulic and one electric power source. The latter three options also
Similar actuation options also exist for the landing gear retraction/extension
found on in-service aircraft. Electric nose-wheel steering system solutions may also
Other than the actuation functions, multiple options also exist for other aircraft
subsystems. For example, the Environmental Control System (ECS) may use
pneumatic or electric power for cabin pressurization. Depending on the type of power
68
consumed, the Wing Ice Protection System (WIPS) and the Cowl Ice Protection
System (CIPS) may be classified as being pneumatic (P) or electrothermal (E) in
nature (other possibilities such as electroimpulse systems also exist, but are not
the protected surfaces, these systems may be further classified as evaporative (Ev),
running-wet (RW), where the heat supplied is only sufficient for maintaining the
protected surface just above freezing. Yet another classification is possible based on
the mode of operation, as the systems may be of the anti-icing (AI) type, which
operate continuously to prevent any ice formation, or the de-icing (DI) type, which
In certain cases, a subsystem may be present within one architecture but absent
from others. A classic example would be an Electric Taxiing System (ETS), where
the aircraft is taxied on the ground using electric motors that drive the wheels. For
the conventional solution, where the aircraft is taxied using the thrust of the main
may be constructed, as shown in Table 2. Despite the fact that only some major
aircraft subsystems and a limited number of options for each such subsystem were
considered in the MoA, it is clear from the total number of combinations that a
very large combinatorial problem has been developed. The number of combinations
would be even greater if the Hydraulic Power Generation and Distribution System
(HPGDS) and the Electric Power Generation and Distribution System (EPGDS) were
considered as additional rows in the MoA, with nominal system pressures and voltages
within such a large combinatorial space quickly becomes infeasible. To put this
69
in perspective, assuming continuous computer operation and a highly optimistic
evaluation time of five seconds per architecture, exhaustive exploration of even the
partial MoA of Table 2 would take approximately 2.13 years. In such cases, it
out combinations with incompatible technologies, thus reducing the size of the
smaller (and thus more tractable) subset of feasible combinations from the MoA.
Such a TCM is not readily applicable as a filter for the case of aircraft subsystem
architectures. Inspection of the MoA of Table 2 does not reveal any obvious physical
incompatibilities. For example, the presence of an Electric Taxiing System does not
directly preclude the use of hydraulic actuation for any of the actuation functions.
Further, even the use of a particular actuation technology for one actuation function
(e.g., elevator actuation) does not directly preclude the use of an alternative actuation
solution for a different actuation function (e.g., rudder actuation). For instance,
each of the four elevator panels of the Airbus A380 is actuated using a conventional
hydraulic actuator in parallel with an EHA, while each of the two rudder panels is
actuated by two parallel EBHAs [24]. The use of electrothermal wing ice protection
imposes no direct restriction on the solution used for nacelle ice protection, which
may remain conventional (pneumatic) - an example of which is the Boeing 787 [14].
A first attempt to reduce the size of the combinatorial space may be made
70
since the objective here is to compare More Electric architectures against the
conventional SOTA.
elements to the MoA, the attempt to reduce the size of the architectural space is first
linear hydraulic actuators are used to actuate the ailerons, elevators, rudders, and
spoilers. Therefore, only this type of actuator may be considered for the conventional
actuation architecture (giving no further consideration to architectures with rotary
high-lift devices shows that the LE and TE devices of most commercial aircraft
are centrally actuated using PDUs [66]. Similarly, while both rotary and linear
steering gear mechanisms are found, the kinematics of the linear “push-pull” type
systems are such that there are certain steering angles at which only one of the two
linear actuators is able to provide a force, thus removing fault tolerance through
actuator redundancy [67]. Thus, for the purposes of this investigation, it is sufficient
to represent the conventional SOTA for actuation functions with a single actuation
architecture in which
2. Centralized hydraulic Power Drive Units (PDUs) actuate the high-lift devices
Despite the fact that the MoA does not contain any obvious incompatibilities,
there are however a large number of impractical combinations. In the context of this
dissertation, impractical combinations are those which are unlikely to be pursued
71
actuation functions are electrified with the exception of, say, nose-wheel steering,
which remains hydraulic. From an engineering standpoint, this arrangement would
for the electrified actuation functions, conventional hydraulic pumps and piping would
need to be retained solely for the purpose of powering the hydraulic nose-wheel
more, for which one of the primary drivers is safety [24]. Since hydraulic actuation
has been used in commercial aircraft for several decades, it is regarded as a reliable
and proven concept which presents low technological risk. On the other hand, the
industry experience with electric actuation concepts is much more limited, as a result
of which considerable conservatism is seen regarding their use on commercial aircraft.
Due to the consequences arising from adverse failures, the conservatism is naturally
staged approach where the actuation functions are progressively electrified starting
from the less critical ones, in order to minimize the risk. In such an approach,
electric actuation solutions that have already been proven in flight (e.g., electric
thrust reverser actuation for Airbus A380, electric brake actuation for Boeing 787)
represent the lowest risk. These are followed by actuation functions for which some
hardware development and validation have already been undertaken (e.g., nose-wheel
steering [68]). For the flight control surfaces, electrification may commence with
surfaces whose operation is not critical for the continued safety of the flight (e.g. the
high-lift system [69]), in which case the failure probability requirements in terms of
failures per flight hour are less stringent. This may be followed by control surfaces
72
such as spoilers that have surface redundancy [27, 26], where the presence of multiple
spoiler panels reduces the criticality of any one panel. There may also be a preference
in terms of the type of electric actuator being used. For example, within the
and electromechanical type (EMA) for the purpose of control surface actuation. It is
generally accepted that the EMA will be lighter than the EHA designed to the same
actuation requirements. However, the EMA also presents some risk of single-point
jamming failures [70] which have caused many to question its suitability for actuating
the primary flight control surfaces in particular. Even for EHAs, their introduction to
actuators (e.g., Airbus A380 [24]). Given their flight criticality and the relatively
low industry experience with electric actuators, the primary flight control surfaces
may in fact be the last actuation functions on the aircraft to be transitioned fully to
actuation combinations existing within the MoA in which the above rules are
73
Table 3: Packaged electrification of actuation functions (Electrification is indicated
either by ✓or by the name of the type of actuator(s) employed. Abbreviations - EHA:
electrohydrostatic actuator, EMA: electromechanical actuator, H/EHA: hydraulic
actuator & EHA in parallel, HLD: high-lift devices, THSA: trimmable horizontal
stabilizer actuator, Sp.: spoilers, Prim.: primary flight control surfaces)
architecture (all hydraulic). In Package 1, the thrust reversers (TRAS) and wheel
brakes (WBS) are electrified, and in Package 2, the landing gear extension/retraction
EHA-driven spoilers and electrically actuated flaps are added, while in Package 4,
the spoilers are transitioned to EMAs and additionally the THS actuation function
actuators for the primary flight control surfaces, while in Package 6, the conventional
hydraulic actuators are removed completely, yielding an All Electric actuation
architecture. Finally, in Package 7, EMAs are considered even for the primary flight
control surfaces.
Moving next to the solutions considered for the ice protection systems, it may be
noted that pneumatic de-icing systems (P-DI) are not commonly found on in-service
74
Table 4: Reduced combinatorial space due to the application of engineering judgment
Subsystem Options #
Actuation functions Packages P0, . . ., P7 8
Wing ice protection P-Ev-AI, P-RW-AI, E-RW-AI, E-RW-DI 4
Cowl ice protection P-Ev-AI, P-RW-AI, E-EV-AI, E-RW-AI 4
Environmental control system Pneumatic, electric 2
Electric taxiing system Present, absent 2
Total combinations: 512
by increased systems complexity and weight [71]. Further, the use of a de-icing system
in general for nacelle ice protection raises concerns over ice being expelled into the
engine [71]. Also, it has been established that electrical power requirements for fully
evaporative wing anti-icing systems (E-Ev-AI) are prohibitive [71], especially when
the surface area to be protected increases. Thus, the architectural design sub-space for
systems for WIPS (E-RW-AI, E-RW-DI), and only anti-icing electrothermal systems
for CIPS (E-Ev-AI, E-RW-AI), in addition to the conventional evaporative and
The application of the filters described above reduces the size of the architectural
architectures within this sub-space were evaluated for this dissertation. A convenient
and compact way to describe a subsystem architecture from within the possibilities
shown in Table 4 is through a candidate subsystem architecture descriptor, which is
Thus, the baseline architectures for the three vehicle sizes considered are as follows:
75
Figure 9: Definition of Candidate Subsystem Architecture Descriptor
The information input to the subsystem sizing and analysis module includes:
information about the aircraft mission profile, mass breakdown, and geometry
This information is used to first derive the requirements for each of the
power consuming subsystems (the distinction between power consuming and power
generation and distribution subsystems was made in §3.1), which are used to size
major components within those subsystems. As far as propagation to subsequent
analysis modules is concerned, the main parameters of interest are the subsystem
mass, peak secondary power requirements, secondary power requirements over the
course of a mission, and direct drag increments (if any). Detailed technical
76
descriptions of the sizing and analysis approach used for each power consuming
subsystem considered are provided in Chapter 4.
Prior to the sizing of the major elements of the power generation and distribution
The heuristic rules used by the algorithm are in turn derived from the inspection of
Tree Analysis (FTA) or Failure Modes and Effects Analysis (FMEA). With the
of the power distribution elements (pipes, ducts, and cables) required to physically
establish those connections are estimated from a 3-dimensional geometric model of the
aircraft. The mass per unit length for these distribution elements is determined based
on the magnitude of the hydraulic fluid, pneumatic mass flow, or electrical current
flowing through them. The required capacities of the power sources (e.g., hydraulic
pumps and electric generators) are determined based on the evaluation of required
the PGDS are presented in Chapter 5. The major outputs of the subsystems sizing
and evaluation module are the following parameters for each subsystem (including
1. The mass of the subsystem, which comprises the mass of major components
present within the defined control volume of the subsystem. For each subsystem,
77
2. The shaft-power requirements of the subsystem (if any) as a function of time
3. The bleed air requirements of the subsystem (if any) as a function of time
4. The direct drag increments on account of the subsystem (if any) as a function
of mission time
performance analysis to compute the updated mission fuel consumption and vehicle
takeoff gross weight. This in turn sets the stage for the re-sizing of the vehicle in
(iii) simplifying assumptions that were made during modeling or analysis. Therefore,
significantly affected by the epistemic uncertainty, and then analyzing the variation
relevant for the case of the MEA architectures, since assumptions regarding the
Therefore, it is also necessary to determine the impact that changes in the assumed
SOTA have on the performance of the MEA architectures. Since many of the
78
electrical/electronic components within a MEA architecture are currently on a steep
development curve, an alternative way to view this is as an attempt to determine
the projected performance of such MEA architectures under the assumption that
technological SOTA progresses to a certain level. From the point of view of modeling,
these two scenarios are however identical, and is done by strategically associating
The subsystems affect a mission-level metric such as fuel consumption through their
(i) mass, (ii) shaft-power requirement, (iii) bleed air requirement, and (iv) direct
drag increments, as shown in Fig. 10. With the exception of the penalty due to mass,
the remaining three penalties may or may not exist. Changes in subsystem mass
and direct drag increments result in a change in the thrust required. On the other
hand, changes in secondary power off-takes result in a change in engine TSFC. The
∆ẇf (t) = ∆ẇf,w (t) + ∆ẇf,spx (t) + ∆ẇf,bx (t) + ∆ẇf,d (t), (14)
in which the quantities on the RHS are respectively the incremental fuel flow rates
due to the effect of weight (mass), shaft-power extraction, bleed air extraction, and
direct drag increment. When the incremental fuel flow rate is integrated over the
course of a mission, the incremental fuel consumption may be obtained as
Z tf Z tf Z tf Z tf
∆wf = ∆ẇf,w (t)dt + ∆ẇf,spx (t)dt + ∆ẇf,bx (t)dt + ∆ẇf,d (t)dt
ti ti ti ti
in which the quantities on the RHS this time are respectively the incremental fuel
79
Figure 10: Contribution of subsystem effects towards mission fuel burn
effect of the change in subsystems weight ∆wss first, and then the remaining three
components separately. To obtain some analytical insight on the effect of the overall
change in subsystem mass on the fuel burn component ∆wf,w for a single flight phase,
in which R is the distance flown at speed V and lift-to-drag ratio L/D. It is possible
to obtain the component ∆wf,w by breaking up the entire mission into phases or
segments and applying the above relationship successively to each segment, starting
Fortunately, most aircraft sizing and performance analysis tools are able to directly
estimate the effect of changing systems mass on mission fuel burn if the parameters
of the mission and the magnitude of the change in mass are specified as input. If the
flight performance analysis uses a point-mass approximation, then for the purpose
of this estimation it makes no difference whether the total change in weight ∆wss is
input as a direct increment to the OEW of the aircraft, or whether the constituents
80
Table 5: Association of subsystem weights with FLOPS weight override parameters
(namelist WTIN)
of ∆wss (weight changes of individual subsystems) are separately factored into the
weight breakdown of the aircraft. The latter approach is followed in this dissertation
weight breakdown of the vehicle. To this end, the weights of individual subsystems
as computed by the subsystems sizing module are associated with FLOPS weight
divide the flight into main phases or segments (e.g., takeoff, climb, cruise, descent,
landing), and assess these effects for each such segment, assuming no variation of
shaft-power, bleed, drag, velocity, lift-to-drag ratio, and engine TSFC over the course
particular the assumption that the off-takes do not vary over the course of a segment.
electric ECS for pressurization increases with increasing altitude, which implies that
it increases during the climb and decreases during the descent. Similarly, the IPS
81
may be active for only the parts of the climb/descent that are under approximately
22,000 ft altitude, even though those segments themselves may end at or begin from
Such assumptions may be removed through an approach where the entire flight
is divided into numerous small intervals, and the cumulative subsystem effects are
assessed through a backward summation from landing to takeoff [72]. Since the
power off-takes and drag increments are in general not constant over the course of a
flight segment, such a backward-summation scheme is followed in this dissertation.
The approach is to first compute the effect of the net subsystems weight change ∆wss
on fuel consumption (as described above), and then use the following update scheme
to compute the additional fuel consumption due to power off-takes and drag:
(k+1)
(k) ∆W (k+1) ∆Wf
∆CL = = (17a)
1
ρV 2 (k) S 1
ρV 2 (k) S
2 w 2 w
82
The ∆-quantities in this scheme are computed with respect to the corresponding
quantities at the same instant in time for a mission in which the effect of subsystem
weight changes has been accounted for (but not the effect of shaft-power and bleed
and parameters corresponding to it are marked with subscript (·)0 . The flight profile
for the basic mission may be obtained from the aircraft sizing and performance
evaluation tool. While the flight profile may change due to the subsystem weight
changes, it is assumed that there is no further change as a result of the effects
of secondary power extraction and direct drag. Each sub-equation in the update
scheme is described briefly, for an arbitrary time in the mission marked with time
(k)
(a) Change in lift coefficient ∆CL : This is due to the change in vehicle weight ∆W
relative to the same point in time of the basic mission. In this case, the change is
simply due to the mass of additional fuel that will be consumed due to subsystem
secondary power off-takes or direct drag between time index ‘k’ and the end of
the mission
(k)
(b) Change in induced drag coefficient ∆CDi : This is due to the change in the lift
drag polar of the vehicle with the basic mission lift coefficient for the same time
(k) (k) (k)
instant CL,0 and the updated lift coefficient CL,0 + ∆CL
(k)
(c) Change in dimensional drag dependent on lift ∆Dddl : Corresponding to the
(k)
(d) Change in dimensional drag independent of lift ∆Ddil : This is due to the direct
drag increments of the subsystems which may arise from the admission of ram
83
(e) Change in dimensional drag ∆D (k) : Summation of the two dimensional drag
components described above
(f) Change in required thrust ∆T (k) : In addition to the direct contribution due to
to the specific energy rate of the incremental weight ∆W (k) . The thrust change
is obtained by perturbing the thrust balance equation T = D + W Vḣ + V̇g , in
which the specific energy rate Vḣ + V̇g is assumed to be the same as that for the
basic mission at the same instant in time
(g) Updated thrust requirement T (k) : Obtained by adding the thrust change ∆T (k)
(k)
computed above to the thrust T0 of the basic mission at the same instant
(k)
(h) Change in fuel flow rate ∆ẇf : Computed as the difference of the fuel flow
rate obtained by accounting for the updated thrust, shaft-power, and bleed air
off-takes and that obtained with the basic mission thrust and no off-takes. Two
(i) Change in vehicle weight ∆W (k) : Obtained by adding the fuel consumed in the
time interval ∆t(k) to the vehicle weight W (k+1) at time index ‘k+1’
The time index is then decremented to ‘k−1’ and the system of equations is
re-evaluated for that time index. The boundary condition is that the vehicle weight
at the final time index ‘n’ is the same as that for the basic mission, i.e., ∆W (n) = 0.
The fuel consumption due to the effect of shaft-power off-take, bleed air off-take, and
direct drag increments is then the difference in vehicle weight at the first time index
The above equation system can be used to determine the TFI (Eq. 9) of individual
84
By feeding the time histories of the subsystem’s shaft-power requirement, bleed air
requirement, and direct drag increment one at a time through the system of Eq. 17,
(i)
the impact of these on fuel burn can be determined as well, in other words ∆wf,spx ,
(i) (i)
∆wf,bx , and ∆wf,d . Due to the nonlinearities present in the update system of Eq.
(i) (i)
17, the sum of the three individually-computed fuel penalties ∆wf,spx , ∆wf,bx , and
(i) (i)
∆wf,d will in general be (slightly) less than the penalty ∆wf,spx+bx+d computed by
considering the power off-takes and drag simultaneously. In this case, the residual
(i) (i) (i) (i)
quantity R(i) = ∆wf,spx+bx+d − ∆wf,spx − ∆wf,bx − ∆wf,d is simply allocated (added)
to the shaft-power, bleed, and drag fuel penalties in the ratio of their magnitudes.
As shown in Eq. 17h, the fuel flow rate is a function of the flight condition (Mach
number, altitude, and thrust setting) and the shaft-power and bleed air extraction:
The fuel flow rate may be obtained in the manner above by directly incorporating
run multiple times per subsystem architecture, this may not be feasible due to
computational expense. Instead, the higher-fidelity tool may be evaluated off-line
to determine coefficients that relate the fuel consumption penalty to the secondary
power off-takes and the flight condition. Within the mission performance simulation
(on-line), polynomial interpolation may be used to determine the fuel flow rate penalty
corresponding to a given flight and operating condition [39]. In this dissertation, two
differing approaches are used for the characterization of the fuel consumption penalty
due to off-takes. The first uses simple empirical correlations that do not directly
account for the effect of flight condition or engine cycle parameters. The second uses
85
are scheduled with respect to normalized mission time (this is discussed subsequently
in §7.2). In either case, it is convenient to represent the engine’s fuel flow rate as the
product of the thrust-specific fuel consumption (SF C) and the thrust, and account
for changes to these quantities (as shown in Eq. 17) caused by secondary power
off-takes and direct drag increments. Towards this end, the following three parameters
are defined as the ratio of the change in the corresponding quantity brought about
by off-takes and drag increments to the value of the quantity accounting for only
propulsive power requirements:
∆ẇf ∆SF C ∆T
λf f = , λsf c = , λT = (19)
ẇf,0 SF C0 T0
Proceeding from the relationship between thrust, fuel flow rate, and thrust-specific
fuel consumption, ẇf = T · SF C, and noting that it yields for the basic mission
Evaluation of the RHS quantities for each step in the backward-evaluation of Eq. 17
allows the increase in fuel flow rate relative to the basic mission to be obtained. The
two approaches mentioned above differ in the method of evaluation of the quantity
λsf c . The first method is introduced in Chapter 6 and the second in §7.2.
In cases where a reduction in either the fuel consumption or the vehicle’s empty
weight (or both) are predicted for an alternate subsystems architecture, there may
86
be a reduction in the vehicle’s takeoff gross weight (or ramp weight) for the design
mission considered. There are two distinct ways in which designers may respond to
such a scenario:
1. Increase the payload and/or range capability of the vehicle by adding payload
the baseline
It is clear that in the first case, the payload-range capabilities of the two vehicles
are dissimilar, while in the second they are held constant. This dissertation assumes
that the payload-range capabilities are specified as part of the design requirements
of the airplane, and that there is no additional credit given to a design that exceeds
those stated requirements. Therefore, the second option (re-sizing of the vehicle) is
Based on the definition of wing loading WT O /Sw , the following re-sizing rule is
Similarly, based on the definition of thrust-to-weight ratio TSL /WT O , the following
re-sizing rule is derived for the rated sea-level static thrust TSL :
TSL TSL TSL
TSL = WT O =⇒ ∆TSL = ∆WT O + ∆ WT O (22)
WT O WT O WT O
With the incorporation of a relaxation parameter α < 1 into the thrust update
equation for reasons of numerical stability, the two main re-sizing rules may be
summarized as:
∆WT O WT O WT O
∆Sw = W − ∆
( T O/Sw ) (WT O/Sw )2 Sw
TSL TSL
∆TSL = ∆WT O + α ∆ WT O (23)
WT O WT O
87
In addition to the main re-sizing rules presented above, additional rules are enforced
for the planform areas of the horizontal and vertical stabilizers, so as to maintain the
same horizontal tail volume ratio and vertical tail volume ratio as the baseline.
The iterations continue until the error terms ∆WT O , ∆(WT O /Sw ), and
∆(TSL /WT O ) are all within a specified convergence tolerance. The second terms in the
right-hand sides of each of the above expressions allow the wing area and thrust to be
updated not just to guarantee convergence of the takeoff gross weight (WT O → 0), but
also to guarantee that the mathematical system converges to the target wing loading
(∆(WT O /Sw ) → 0) and thrust-to-weight ratio (∆(TSL /WT O ) → 0). In general, the
target wing loading and thrust-to-weight ratio may be updated in between iterations
based on a re-evaluation of the constraint analysis (Eq. 11 and Fig. 8). The targets
may change due to changes in the weight fractions β or the drag polar coefficients
would be small enough that the original feasible design point would continue to be
feasible, with perhaps a slight alteration in the margin of the design point from
the active constraints. Pursuant to this assumption, a constant wing loading and
seen from Fig. 11, changes in the mass of other subsystems due to changes in the
subsystem whose architecture has not been altered (e.g., FCAS, as shown in Fig. 11).
rules of Eq. 23 that this will precipitate a scaling down of the wing planform area
88
Figure 11: An example of establishment of dependencies among subsystem-, aircraft-,
and mission-level parameters: Due to re-sizing, mass changes originating in other
subsystems result in changes in lifting surface planform area, which change actuation
loads and thus the mass of the surface controls sized to those loads. Thus, the surface
controls mass may change even without architecture changes to the FCAS itself.
and the rated thrust, both of which will tend to reduce the ramp weight further (an
amplifying effect). For the same reason, if a particular subsystems architecture results
in the ramp weight of the aircraft increasing, then re-sizing will cause a scaling up
of both the wing planform area and the rated thrust, which will in turn will tend
to further increase the aircraft weight. It may be argued that it would be pointless
to attempt re-sizing in this second case, given the amplifying effect. However, the
1. For the case of MEA architectures, it is quite possible that the empty weight
of the aircraft increases while the fuel consumption reduces. The relative
the vehicle’s ramp weight. Despite this, the architecture is still of interest on
2. Such comparisons are more appropriate when the designs which have the same
thrust-to-weight ratio and wing loading, since these essentially imply that they
89
Based on the above arguments, re-sizing of the aircraft is carried out regardless of
whether the subsystems architecture results in an increase or decrease in the vehicle’s
ramp weight. Therefore, all comparisons between subsystem architectures that are
presented within this dissertation are between aircraft that have both the same
For each run of the integrated sizing and analysis environment, information
generated regarding the sized aircraft’s geometry, subsystem architecture, and mission
subsequently be run on the generated data. Those presented in this dissertation are
mainly of the following types:
This chapter introduced the subsystems that are considered in this dissertation
further subsequently. The major elements of the integrated subsystem sizing and
analysis method such as the generation of candidate architectures, the evaluation and
90
decomposition of subsystem impacts on aircraft performance, and the simultaneous
re-sizing of the aircraft and subsystems were presented and discussed. While the
and distribution subsystems, the sizing and analysis approach for subsystems falling
within these two categories was not presented in this chapter to preserve continuity
where the modeling approach for each of the subsystems is presented in detail.
91
CHAPTER IV
The power consuming subsystems are those that are the end-users or consumers
of secondary power which is transmitted by the appropriate power generation and
distribution systems (the latter are covered in Chapter 5). In the scope of this
The remainder of this chapter is organized as follows: For each power consuming
subsystem, the control volume for analysis is first defined. Subsequently a brief
description of the function and major components is provided, followed by a discussion
on applicable tradeoffs associated with novel solutions for the subsystem. The
modeling and simulation approach for each subsystem, whose end goal is to estimate
the mass, secondary power requirements, and drag increments, is then described.
92
4.1 Flight Controls Actuation System (FCAS)
The FCAS control volume includes the surface controls for all the control surfaces of
the aircraft. Depending on the type of actuation being employed, the power input to
this subsystem may be hydraulic, electric, or a combination of hydraulic and electric.
The output power is always in mechanical form (movement of the control surfaces).
Therefore, the power off-take from the engine in order to satisfy the requirements of
this subsystem is always in the form of shaft-power. The subsystem does not utilize
any bleed air and does not result in any direct drag increment since it is within the
hydraulic actuators or power drive units (PDUs). However, there is a recent trend
towards the use of electric actuators either in parallel with conventional hydraulic
actuators and serving a backup role (e.g., Airbus A380 [24]) or for the actuation of
control surfaces such as spoilers which have surface redundancy (e.g., Boeing 787).
electric actuator designed for the same application [73]. Since the tradeoff must also
account for weight reduction of the hydraulic system and weight addition to the
electric system, it may not be necessary for the electric actuator to be lighter than
hydraulic actuator it is replacing also exists for actuation functions other than flight
control surface actuation. In each of these cases, the benefit has to be ascertained
93
4.1.1 Control Surface Descriptions and Layouts
The purpose of the flight control surfaces is to ensure adequate controllability and
maneuverability for the aircraft and provide the required flight characteristics for a
given mode of flight. While the number and layout of control surfaces varies greatly
depending on the size and type of aircraft, the following characteristics are in general
• Primary flight control surfaces: The primary flight control surfaces are the
ailerons, the elevators, and the rudder. These are hinged bi-directional
trailing-edge control surfaces that are responsible for control of the aircraft in the
roll, pitch, and yaw axes respectively. For small aircraft, it is common to employ
two ailerons (one per wing), two elevators, and one rudder. Larger aircraft
may employ four elevator panels, two rudder panels, and up to six aileron
panels (three per wing). These control surfaces have high control authority,
• Spoilers: Multiple spoiler panels are located on the upper surface of the wing.
Flight spoilers may be deflected symmetrically on both wings as speed-brakes
steeper descent. Further, on almost all modern commercial aircraft, the roll
touch-down to aid in decelerating the aircraft. For this purpose, aircraft may
also have dedicated ground spoilers, which are designed to only deploy upon
spoiler droop function as well [74]. Due to the presence of multiple spoiler
94
panels, surface redundancy already exists, and no single spoiler panel is
flight-critical. As a result, spoilers are actuated by only one actuator per panel.
satisfy the required approach speed and field performance constraints without
large planform area wing. The types of high-lift devices used varies greatly
between aircraft of different sizes and between manufacturers, and an excellent
contains slats and/or Krueger flaps that cover nearly the entire wingspan
except for the mounting positions of the engine pylons. Krueger flaps have
high stowing loads and may also see actuation load reversals, and modern
significant complexity and weight penalties, and in modern aircraft are avoided
the entire horizontal stabilizer can pivot about an axis. Due to its location and
planform area, the THS has extremely high pitch control authority and is made a
of flaps and/or landing gear), trim is re-established using the elevator. Following
this, the THS may be moved to a new setting that allows the established trim
95
Table 6: Summary of control surfaces for SSA, LTA, and VLA baselines
The aerodynamic loads seen by the control surfaces are directly affected by their
size (planform area), which can only be finalized after detailed dynamic analyses of
control effectiveness, structural loads, and control system effects. Certain guidelines
are nevertheless available for initial sizing [37]. The empennage surfaces (elevator
and rudder) typically begin at the side of the fuselage and extend to the tip of the
corresponding stabilizer (or at least to 90-95 % span). These control surfaces are
typically tapered with the same chord ratio as the main lifting surface, allowing for
a straight-tapered spar and yielding a constant chord fraction. For jet transports,
Raymer [37] gives a chord fraction of 25 % for the elevators and 32 % for the rudders.
For ailerons, Raymer [37] provides historical guidelines that relate the flap chord
ratio of these surfaces to their spanwise extension. Depending on whether the slats
are constant-chord or tapered, the chord ratio relative to the local wing chord may
vary between 6.7 % to 16 % at the root and between 16.5 % to 33 % at the tip [66].
For trailing-edge devices, the ratio of flap chord to local wing chord in the stowed
The control surface sizes established using the above guidelines were cross-verified
by reference to existing commercial aircraft similar to the SSA, LTA, and VLA. The
control surface layouts and actuator positions are summarized in Table 6 and Fig. 12.
The parametric definition of control surfaces with respect to parent lifting surfaces
allows the former to re-size if aircraft re-sizing changes the geometry of the latter.
96
Figure 12: Control surface definitions for SSA, LTA, and VLA baselines
97
4.1.2 Actuation Loads for Ailerons, Elevators, and Rudders
bm /2
yo b2m
yi AR = Sm
yi
ηi = bm /2
λc/4
yo
Sm /2 ηo = bm /2
c̄ c̄f =
cf
c̄
λhl cf ch ch
Sf c̄h = cf
Figure 13: Control surface geometry for hinge moment coefficient estimation
For control surfaces hinged to the trailing edges of lifting surfaces (i.e, ailerons,
that comprises a baseline contribution due to control surface camber (if any), a
contribution due to local effective incidence angle (αef f ), and a contribution due
to control surface deflection (δ),
Ma = q̄Sf cf Ch
where q̄ is the dynamic pressure, and Sf and cf are respectively the planform area and
the chord of the control surface. A slight manipulation of Eq. 24 yields the following:
= [Ch,0, Ch,α (M), Ch,δ (M, δ)] . [1, αef f , δ]T q̄ . Sf c f (25)
This requires information such as the planform area (Sf ), the mean chord of the
control surface (cf ), flap chord ratio (c̄f ), airfoil shape, hinge-line sweep (λhl ), main
surface sweep (λc/4 ), spanwise coordinates of the inboard and outboard extremities of
98
the surface (ηi , ηo ), and distance between the hinge-line and the leading edge of the
control surface (c̄h ) (Fig. 13) to generate a 1-dimensional lookup table for Ch,α (M) and
a 2-dimensional one for Ch,δ (M, δ). This information is automatically extracted from
the 3-D geometric model. In previous work [76], the hinge moment estimates obtained
from the NASA F-18 SRA program [20], as shown in Fig. 14. The agreement between
10 10
0 0
−10 −10
−20 −20
0 1 2 3 4 5 0 1 2 3 4 5 6
Time (s) Time (s)
4000
Hinge Moment (Nm)
2000 Predicted
1000 Flight Test 2000
0 0
−1000 Predicted
−2000
−2000 Flight Test
−4000
0 1 2 3 4 5 0 1 2 3 4 5 6
Time (s) Time (s)
(a) Abrupt roll doublet - full stick (left): M (b) Abrupt aileron reversal - full stick (left):
= 0.72, h = 25 kft, q-bar = 13.60 kPa (284 M = 0.84, h = 25 kft, q-bar = 18.29 kPa (382
lbf/sq.ft.) lbf/sq.ft.)
Figure 14: Comparison of predicted hinge moments with NASA F-18 SRA flight test
data (same geometry, flight condition, and surface deflections as input)
The parameters in the second group of Eq. 25 collectively define a flight condition.
The maximum hinge moment corresponds to a flight condition that maximizes the
product of the first two groups. Given the Mach number and deflection angle
dependencies of the coefficients of the first group, finding this critical flight condition
requires the exploration of multiple points in the aircraft altitude-speed flight envelope
in search of the peak hinge moment. The guidelines are provided in Part 25 of
the Federal Aviation Regulations [77], and are further discussed in the works of
99
Figure 15: Rudder authority limiting relationship as function of airspeed
• Ailerons (FAR §25.349) - Three flight conditions are considered: (i) Full
deflection at design maneuver speed (VA ), (ii) a deflection generating the same
roll rate as (i) at design cruise speed (VC ), and (iii) a deflection generating 1/3
(shown above), there is not an abundance of handbook methods for the calculation
100
of spoiler hinge moments. Scholz [78] reviewed available data pertaining to spoilers
from literature and also from aircraft manufacturers and suggested the following two
where Ssp and csp are the spoiler planform area and chord, δsp its deflection, c(ysp )
the wing chord at spanwise location of spoiler, b the wingspan, W the aircraft weight,
and nz the load factor. The author recommended CD = 1.5 − 1.8, Vl = 1.14 V∞ , and
Ksp = 1.5, which gave a good match with available spoiler data.
In some aircraft, certain spoiler panels are dedicated to only the ground spoiler
role and are not used in flight. Further, the deflection limits for all the spoilers are
not necessarily identical. However, for simplicity, this dissertation assumes that:
1. All spoiler panels serve both flight spoiler and ground spoiler roles
2. For each spoiler panel, maximum physical deflection limit is 60◦ trailing-edge
up, with maximum in-flight deflection limited to 38◦ (based on [82, 83])
With these assumptions, the following flight conditions are evaluated in order to
in-flight deflection during an emergency descent at the design dive speed (VD )
For all three baseline aircraft, extension in flight to the permissible limit generated
101
4.1.4 Actuation Loads for High-lift Devices
With only a few exceptions, the power required to actuate the high-lift systems
of modern commercial aircraft comes from centrally located Power Drive Units1
(PDU). In a conventional PDU, hydraulic motors (typically two) drive into a gearbox
which converts hydraulic power to mechanical rotary output motion. This motion
or linear actuators which ultimately provide motion to the high lift devices. Such
which is further verified by position pick-up sensors. In case these sensors detect
asymmetry that exceeds a defined threshold, brakes engage to immobilize the drive
straightforward a manner as those for the hinged control surfaces described previously.
The actuation loads are functions of not only the dimensions of the leading-edge
devices (LED) and trailing-edge devices (TED), but also the kinematics of the
associated mechanism. Rudolph [66] provides an excellent technical description of the
high-lift systems of a large number of commercial aircraft, from which it is evident that
there is significant variation in the flap design philosophy (including flap mechanisms)
Given the difficulty of assessing the actuation power of the high-lift system
through a direct assessment of the aerodynamic loads, the actuation power is instead
determined based on the power ratings of the PDUs. Based on PDU information
collected from a number of sources [82, 84, 39] and the corresponding Maximum
Takeoff Mass (MTOM) of each aircraft, the installed PDU power as a function of the
MTOM is seen to follow the trend shown in Fig. 16. This power is typically supplied
1
Power Drive Units (PDUs) are also often referred to as Power Control Units (PCUs)
102
Figure 16: Variation of flap Power Drive Unit (PDU) rating with MTOM
by two power sources (e.g., hydraulic motors) which collectively supply the PDU. In
the event of failure of one power source, the other power source can continue to drive
the flap system, but at a reduced rate. For simplicity, it is assumed that both the
TEDs and the LEDs are actuated by PDUs of identical design, upon verification that
this is, in fact, the case for certain in-service aircraft [82, 84].
The hinge moment estimation technique used for the aileron, elevator, and rudder
is also applicable to the THS. However, while the hinge moment coefficients can be
which the THS will see the maximum hinge moment. For modern FBW aircraft,
the THS under normal operating mode is automatically manipulated by the FBW
system to allow the elevator to be offloaded and returned to a faired position following
the establishment of pitch trim. This manipulation logic determines the amount of
mis-trim that is permitted before the THS is actuated and therefore directly influences
the magnitude of the hinge moments developed. Since this logic is determined much
the actuation requirements of the THS and the subsequent mass estimation for the
103
Figure 17: Variation of Trimmable Horizontal Stabilizer Actuator (THSA)
characteristics with MTOM
Information regarding the THSA mass, maximum load F0,thsa , and stroke sthsa for
a range of Airbus aircraft from 1970 to the present date was provided by Socheleau
et al. [85]. Using this information and the publicly available information regarding
the MTOM for each of these aircraft, the trends shown in Fig. 17 are obtained.
The stop-to-stop range of motion of the THSA is similar for most commercial
104
slow-moving control by design, with a maximum rate around δ̇max = 1◦ /s. This
allows the maximum rate of the THSA to be estimated as
sthsa δ̇max
vmax,thsa = . (27)
∆δmax
cp
Pthsa = F0,thsa vmax,thsa (28)
where Mdrive is the mass of the drive system, which for a conventional THSA typically
comprises two hydraulic motors. However, in recent aircraft such as the Boeing 787
the THSA has been electrified [74], in which case electric motors are used. For either
case, the component Mdrive is estimated using the generic actuator mass estimation
technique described in the following section.2
Mass estimation of electric actuators was addressed in Chakraborty et al. [86, 87]
and Garmendia et al. [88, 89] using a component mass build-up approach, in which
the masses of the major actuator components were evaluated based on the output
requirements (load, force, and stroke) and subsequently summed to yield the actuator
mass. A limited number of actuator and kinematic design variables was used to
search for mass-optimal actuator designs. With regard to the current problem,
one limitation of this approach is that explicit models are required for the main
2
Technically, the mass M0,thsa as obtained from Fig. 17 already accounts for the mass of the
hydraulic motors. However, these motors account for only a very small fraction of the overall THSA
mass. Thus, Eq. 29 provides a reasonable (and conservative) estimate of the THSA mass.
105
focuses more on the variation of actuator mass with technological SOTA than on
mass-optimal actuator designs, a more generalized approach to mass estimation has
been used. For simplicity, this is first developed for the case of a linear electric
actuator and then further generalized to the case of any actuator and output motion.
For an electric actuator with linear output motion, let (F0 /M) define the ratio
of the actuator’s stall load to its mass (units: kN/kg or lbf/lbm). Further, let
(P/M)em and (P/M)pe be the power-to-mass ratios of the electric motor and the
power electronics (units: kW/kg or hp/lb) with current SOTA. These components are
assumed to have overall efficiencies ηem and ηpe respectively. The overall efficiencies of
actuator components downstream of the electric motor output shaft are collectively
where κem = 1 and κpe = 1 indicate current SOTA as far as electric motor and power
electronics technology are concerned. Improvements in the SOTA are represented by
values of κem > 1 and κpe > 1, using which it is clear that a reduction in the actuator
The tacit assumption in this approach is that the design of mechanical components
components such as pumps and cylinders (found in an EHA, for instance) has already
reached a high level of refinement. Thus, for these actuators, substantial mass
106
Table 7: Assumed figures of merit for actuator types and components
where X0 may represent the stall load F0 for a linear actuator (Eq. 30), the maximum
output of a PDU. The corresponding figure of merit (X0 /M) of the actuator may
simply ignoring the second term for the electric motor and power electronics.
peak stall load F0 . If ∆pwk is the maximum hydraulic pressure difference available at
the site of the actuator, then the required actuator area may be estimated as
κ F0
Areq = , (32)
∆pwk
where κ > 1 provides an excess margin. Once the required hydraulic actuator area
has been sized, it is clear that the required flow rate is directly proportional to
the actuator’s rate or output velocity. Thus, if the actuation function requires the
107
generation of an output velocity v, then the corresponding flow rate requirement is
κ F0 v
Q = Areq v = . (33)
∆pwk
This flow rate requirement becomes maximum when v = vmax , i.e., when the
κ F0 vmax κ Pcp
Qmax = Areq vmax = = , (34)
∆pwk ∆pwk
where Pcp = F0 vmax is the corner power of the actuator (product of maximum output
force and maximum output velocity). From the above, the following two observations
1. From Eq. 33, it is seen that the flow rate required is directly proportional to
the required output velocity. For the case of flight control surfaces, this implies
that the maximum load (in terms of required flow rate) is imposed when the
control surfaces move at high rate (angular velocity)
2. From Eq. 34, it is seen that for a given corner power Pcp , the maximum flow
Thus, the same actuation load when served by a higher pressure system (e.g.,
5,000 psi instead of 3,000 psi) will have a lower fluid flow rate requirement
air loads acting on a control surface. Let the control surface hinge moment Mh be
expressed as a function of control surface position δ as
Mh = A + Bδ. (35)
Further, let the mean gearing ratio (Fig. 18) between the control surface and the
108
∆δ
Gk = ∆x
∆x
Airframe
Control surface
r
uato
Act ∆δ
velocity of the electric motor and the maximum speed of the output ram, i.e.,
ωmax ω
Gact = = (38)
vmax v
Gkin δ̇/v δ̇
= = = Gef f (39)
Gact ω/v ω
Assuming an actuator efficiency ηact that accounts for the motor windage losses and
losses in components downstream of the electric motor, the required torque output
Gef f
τm = Mh = κMh (40)
ηkin ηact
Considering only mechanical shaft-power output and heat losses, the required
109
where Km is the motor constant which also relates the torque and the current as
τm = Km I. Using this, the power requirement becomes
R 2
Pin = I 2 R + Km ω I = 2 m
2
τ + τm ω = λ τm + τm ω (42)
Km
Using Eqs. 35, 39, and 40 in Eq. 42, the power requirement of the motor may be
δ̇
Pin (δ, δ̇) = λ {κ(A + Bδ)}2 + κ(A + Bδ)
Gef f
2 2 2 2 2 2 κA B
= λκ A + (λκ B ) δ + (2λκ AB)δ + δ̇ + δ δ̇(43)
Gef f ηkin ηact
Subject to the assumptions stated above, Eq. 43 is valid for any arbitrary control
surface motion. Unlike hydraulic actuation, it is seen that for electric actuation the
For evaluating the power requirements of control surface actuators during the course
where Ω = δ̇peak /δpeak is the ratio of the peak control surface rate to the amplitude of
excursion. The excursion amplitude is bounded by the deflection limits of the control
surface, i.e. 0 ≤ δpeak ≤ δmax . The peak rate is limited to the maximum permissible
and the average power over one sinusoidal cycle period T = 2π/Ω is computed, then
110
• Term 1: This propagates unchanged, and represents the electrical power
required to provide the holding load to maintain the surface at δ = 0
√
• Term 2: The average value of this term is obtained by setting δ = δmax / 2
• Term 3: The average value of this term over the time period is zero
• Term 4: The average value of this term over the time period is zero
zero. Physically this is due to the fact that the power recovered (regenerated)
during motion with aiding load (Mh < 0, δ̇ > 0 and Mh > 0, δ̇ < 0) is
load (Mh > 0, δ̇ > 0 and Mh < 0, δ̇ < 0). However, if the re-introduction
of regenerated power onto the electric bus is disallowed (by assuming that
regenerated power is dissipated through a shunt resistor), then for this term
the average value of the power supplied over the cycle may be obtained by
setting δ δ̇ = δmax δ̇max /(2π)
Incorporating these observations, the mean input power requirement of the motor
R Gef f
λ= 2
, κ= . (46)
Km ηkin ηact
Finally, if the prescribed motion only occurs for ǫ ∈ [0, 1] fraction of the overall
time, and if the efficiency of the actuator’s power electronics is ηpe , then the mean
111
Applying the prescribed sinusoidal motion for the case of the hydraulic actuator,
the mean hydraulic power may be estimated as
!
2 ǫ δ̇peak
P̄hyd = Pcp , (48)
π ηact δ̇max
where ηact is the assumed overall efficiency of the hydraulic actuator. The quantity
2/π appears as the absolute value of the angular rate is integrated over a cycle.
Within the scope of this dissertation, the control volume of the LGAS includes
only the retraction/extension actuator. Additional actuators such as the uplock and
downlock actuators and door actuators are not modeled. Depending on the actuator
type, hydraulic or electric power may be required. Therefore, this subsystem always
actuation load. Of these, the effect of the gravitational load (due to the mass of
the landing gear leg) is dominant [95]. An analysis of various aircraft showed that
the total landing gear mass is approximately 3 % of the total mass of the airplane,
with the nose gear accounting for 0.5 % [96] (in other words around 16.7 % of the
total landing gear mass). However, a more detailed component mass breakdown for a
number of transport aircraft of varying sizes [97] shows the contribution from the nose
landing gear to the total landing gear mass. From the mass breakdown presented in
[97], the relative location of the CG of the main landing gear (ℓ̃cg,mlg,0) and the nose
landing gear (ℓ̃cg,nlg,0) were also computed. These are shown in Table 8 for the three
112
Table 8: Relative CG position of main landing gear leg (ℓ̃cg,mlg,0) and nose landing
gear leg (ℓ̃cg,nlg,0) for SSA, LTA, and VLA baselines
baselines. The quantities are normalized using the total length of the gear legs (Lmlg
and Lnlg ) from the pivot point to the ground contact point, thus yielding a relative
location of 0 for the pivot point and 1 for the ground contact point.
Since most aircraft sizing tools (including FLOPS) provide an estimate of the
total landing gear group mass Mlgg , this is used with the percentage contributions
discussed above to compute the mass of the nose and main landing gear legs as
where nmlg is the total number of main landing gear legs. Thus, the simplifying
assumption was made that for aircraft with nmlg > 2 (e.g., presence of two
wing-mounted main gears and two body-mounted main gears), the masses of the
main landing gear legs are identical. The approach to computing the landing gear
actuation requirements is identical for nose and main landing gears, and subscripts
The mass and relative CG location of the landing gear legs is affected by the mass
and locations of subsystems such as the WBS, NWSS, and ETS (if installed). The
subscripts ‘0’ in Eq. 49 and Table 8 signify that the corresponding quantities apply
to the baseline vehicle with conventional subsystem architecture (hydraulic WBS and
NWSS, no ETS installed). Alternative actuation solutions for the WBS and NWSS
may result in either an increase or decrease in the mass of the landing gear leg. The
incorporation of ETS on the other hand always results in an increase of mass of the
113
Figure 19: Landing gear retraction kinematics
leg. The net mass and effective CG position for the landing gear leg considering the
where ∆Mss,i is the change in mass of the ‘i’th subsystem relative to the conventional
architecture, and ℓ̃cg,ss,i is the relative location of subsystem ‘i’ on the gear leg.
To relate the mass properties of the landing gear to corresponding actuation
requirements, it is assumed that the main landing gear legs pivot about an axis
parallel to the aircraft’s longitudinal axis, retracting into the main landing gear bay.
Similarly, the nose gear is assumed to pivot about an axis parallel to the aircraft’s
lateral axis and retract forward into the nose landing gear bay [98]. The effect of
gear mechanism. The same kinematic relationships are also valid for the nose gear.
point, C the landing gear pivot point, and G the landing gear leg center of gravity.
114
Table 9: Main and nose landing gear retraction mechanism parameters
Considering moments about the landing gear pivot point C, the length c (related
to actuator ram position) and the required actuator force F may be computed as a
computed as
s
F0 = max F (θ), s = c(θdn ) − c(θup ), vmax = , (52)
θ∈[θup ,θdn ] ∆t (1 − ǫ)
where θup and θdn correspond to the retracted and extended positions of the gear
respectively, and ∆t is the retraction time. The mechanism is assumed to have
angular acceleration/deceleration over equal time intervals ǫ · ∆t, ǫ ∈ (0, 1). For the
remaining time ∆t (1 − 2ǫ), retraction at the maximum ram speed vmax is assumed.
The settings for the kinematic parameters appearing in Eq. 51 are summarized
in Table 9. They are set based on the description of a commercial transport landing
gear mechanism provided by Young [98]. The dimensions a and b are set as fractions
of the landing gear length. The force-stroke characteristic that is obtained through
Eq. 51 using the main landing gear parameters is shown in Fig. 20, and satisfies the
preliminary requirement that the force be relatively uniform over the majority of the
115
1
0.8
Normalized
0.6
Actuator
Force Note: The following are used
0.4 for normalization −
0
0 0.2 0.4 0.6 0.8 1
Under the assumption that the landing gear is retracted with the actuator operating
at its maximum rate vmax , the corner power and peak power corresponding to the
cp
Plga = F0 vmax ,
peak θup
Plga = max F (θ) vmax .
θ=θdn
For both hydraulic and electromechanical landing gear actuators, the mass is
determined using the generic actuator mass relationship described in Section 4.1.6,
116
4.3 Nose-wheel Steering System (NWSS)
The control volume for the NWSS consists of the steering mechanism, which may use
either hydraulic power or electric power. In either case therefore, the NWS imposes
only a shaft-power requirement. No direct drag is attributed to this subsystem.
Electrification of the NWS function was considered as part of the ELGEAR and
DRESS programs3 [68, 67, 100], in part motivated by the interest in eliminating
flammable hydraulic fluid from close proximity to heated brakes in landing gear bays.
The optimized DRESS actuator was found to be heavier than the baseline hydraulic
The sizing of the steering mechanism requires the estimation of the maximum
required steering moment. The applied moment must overcome the opposing moment
generated due to frictional forces between the tire and the ground. These frictional
forces increase with the load being carried by the nose gear leg. The recommended
range for nose gear static loading is 8-15 % [37, 96]. Below this range, nose-wheel
steering authority may be compromised while above this range, takeoff rotation
limit. Under these conditions, the nose gear static load may be assumed to be ǫs =
15 % of the total aircraft weight. The nose gear load may increase further under
braking conditions due to the load transfer from the main gear to the nose gear.
Therefore, in accordance with FAR §25.499, a dynamic load factor κ = 1.33 is applied
to the static loading to obtain the maximum nose gear reaction to be used for NWSS
3
ELGEAR: Electric Landing Gear Extend and Retract, DRESS: Distributed and Redundant
Electromechanical nose wheel Steering System
117
θr
wtire
Dnom
2
Rroll
Tnw xtire
sizing:
max
Fz,nlg = κ ǫs (MRW ) (54)
Also in accordance with FAR §25.499, a critical horizontal (lateral) load at the
tire-to-ground contact point of 0.8 times the normal reaction of Eq. 54 is considered,
nlg
F
i.e., Fynlg = 0.8. The lateral load developed is thus given by
z
! !
max
Fynlg max
Fynlg
Fy,nlg = Fz,nlg = κ ǫs (MRW ). (55)
Fznlg Fznlg
The moment that is developed about the steering axis as a result of the lateral
max
force Fy,nlg may be determined based certain geometric parameters of the steering
mechanism, which are shown in Fig. 21. The rake angle θr is simply the inclination
of the steering axis to the vertical. This angle and the axle offset result in the
generation of the tire trail xtire , which is the distance between the theoretical ground
contact point and the intersection of the steering axis with the ground, measured
along the ground. The rake angle and trail are set as shown in Table 10 based on
early design guidelines for this category of aircraft [37]. This geometry results in the
118
Table 10: Nose landing gear and NWSS parameters
using which the maximum steering moment for this case may be computed as
!
nlg
nws max
Fy
Msteer = Lef f Fy,nlg = (xtire cos θr ) (κ ǫs MRW ). (57)
Fznlg
In addition to this, the steering moment required in order to steer the nose gear
with the aircraft stationary must be computed. In this case, since there is no forward
motion of the aircraft, a certain degree of tire scrubbing occurs. The friction forces
generated as a result of this once more translate into a moment about the steering
axis which must be overcome by the steering system. An empirical relationship for
nws max b2
Mscrub = 0.24 Fz,nlg (imperial units),
Tnw /2
s 2
δ δ
b = 2wtire − ,
wtire wtire
Dnom
δ = − Rroll , (58)
2
in which the wheel track Tnw , nominal tire diameter Dnom , rolling radius Rroll , and
tire section width wtire are in inches [101]. Dnom , Rroll , and wtire are obtained by the
process of tire selection, which is explained in Section 4.4. The sizing moment for the
NWSS is taken to be the maximum of those predicted by Eq. 57 and Eq. 58. The
steering moment for the SSA obtained using this approach was in good agreement
with those published from the ELGEAR project [68] for the Airbus A320 aircraft.
119
1 1
0.5 0.8
Normalized
Maximum 0.6
0
Moment
0.4
−0.5
0.2
−1
0
−1 −0.5 0 0.5 1 0 0.2 0.4 0.6 0.8 1
Normalized Steering Angle Normalized Rate
Figure 22: Load-angle and load-speed envelopes for nose-wheel steering system
To determine the power requirements for the NWSS, the load-angle and load-rate
rate. The peak power and corner power of the steering mechanism are then given by
δ̇max
peak
Pnws = max δ̇ Mmax (δ̇) ,
δ̇=0
cp max
Pnws = Mnws δ̇max . (59)
Assuming an overall efficiency of ηh−nws and ηe−nws for hydraulic and electric
120
4.3.3 Mass Estimation
For the electric NWSS, the mass of the gearbox is estimated using an empirical
correlation developed at NASA using actual gearbox weight data [102], where gearbox
mass is a function of the maximum delivered output power and the gear ratio:
The horsepower (hp) in Eq. 61 is obtained from the peak power computed in Eq. 59.
The input and output RPM correspond directly to the maximum angular rate of the
NWSS and the maximum RPM of the electric motor (Table 10). The motor and
where (P/M)em and (P/M)pe are the power-to-mass ratios of the electric motor and
the power electronics. ηgb , ηem and ηpe are the overall efficiencies of the gearbox,
motor, and power electronics respectively. The electric NWSS mass is summed as
No model of the hydraulic steering mechanism was created in the scope of this
dissertation. Instead, the mass of the hydraulic steering system is estimated based
on weight breakdown data presented by Currey [97] for the controls associated with
The function of the wheel braking system is to decelerate the aircraft on the ground.
During the deceleration, the kinetic energy of the aircraft is converted to thermal
(heat) energy which is dissipated through the brakes. In a conventional actuation
121
architecture, hydraulic brakes are used. However, as with landing gear actuation
and nose-wheel steering functions, there is significant interest in electrifying brake
actuation as well, since it would allow the elimination of flammable hydraulic fluid
from the vicinity of heated brake disks. Electromechanical brakes are already in
service on the Boeing 787 aircraft [30, 74]. Depending on the aircraft size, the weight
For the sizing of the brakes and brake actuation system, the aircraft deceleration
requirements must be converted into brake actuation requirements. The total braking
force Fbtot between the tire and the ground required to generate a desired deceleration
M V̇ = T − D − Fbtot − Ff + Mg sin θ,
In Eq. 64, M is the aircraft mass, T the total thrust of all operating engines
(T < 0 implies reverse thrust), D the aerodynamic drag, Ff the rolling friction force,
and θ the surface gradient (θ > 0 is the aircraft is moving down-hill).
from nbrk braked wheels, then the braking force per wheel is given by
Fbtot
Fb = . (65)
nbrk
The braking torque per wheel (torque applied by the braking system to the braked
wheel) may be obtained from the braking force per wheel by considering the equation
122
Figure 23: Geometry of brake rotors and stators
In Eq. 66, Rroll is the rolling radius of the braked tire, and the effect of wheel inertia
Iw ω̇w is neglected in comparison to the other term. Modern aircraft brakes consist of
alternate “rotor” and “stator” elements that collectively form a “heat-stack” (Fig. 23).
The rotors are keyed to the wheel and rotate with it, while the stators are keyed to
the hub and are stationary. For N rotors, 2N total friction surfaces are formed. The
braking torque τb and the required axial force Fa may be related as:
Z ro Z ro
τb = (2N) r.p(r).dA, Fa = p(r).dA, dA = 2πr.dr, A = π(ro2 − ri2 )
ri ri
(67)
distribution is assumed, then p(r) = p = Fa /A. On the other hand, if uniform wear
(“u/w”) is assumed, then r·p(r) = constant (this case is assumed in this dissertation).
With these assumptions, the torque and the axial force are related as follows:
2 ro3 − ri3 ro + ri
τb = C(.) Fa , ... Cu/p = (2N) µ , Cu/w = (2N) µ , (68)
3 ro2 − ri2 2
where ro and ri are respectively the outer and inner radii of the friction surfaces and
123
Table 11: Static cases for establishing braking requirements
Case description Full-throttle run-up Parking on incline
Aircraft weight MTOW MRW
Thrust setting Max. SLS -
Ground speed - -
Deceleration - -
◦
Opposing gradient 1.1 1.1◦
The brake actuation requirements are determined from the two static cases (aircraft
stationary) and two dynamic cases (aircraft decelerating) shown in Tables 11 and
12 respectively. The static cases correspond to the ability of the brakes to hold the
aircraft stationary during a full-throttle run-up and while parked on an incline of a
specified gradient. The two dynamic cases are an accelerate-stop case corresponding
For each of the conditions listed in Tables 11 and 12, the total required braking
force Fbtot is evaluated using Eq. 64. For a conservative estimate, factors in this
relationship that reduce the braking requirement, such as longitudinal load transfer,
rolling friction, and the aerodynamic drag are neglected (Ff ≈ 0, D ≈ 0).
However, if the brakes are sized merely to these requirements, they will be
the torque on the wheel due to the friction at the tire-ground interface exceeds the
124
Table 13: Number of main landing gear wheels and braked wheels
MRW # of MLG # of braked
(lb) wheels MLG wheels
[50, 000 − 200, 000) 4 4
[200, 000 − 550, 000) 8 8
[550000 − 820, 000) 12 12
[820, 000 − 1, 100, 000) 16 16
1,100,000 + 20 16
available braking torque [104]. This is not typical of modern commercial aircraft,
which are equipped with Anti-lock Braking Systems (ABS) to prevent the wheels
from locking up under braking. ABS is designed to modulate braking pressure and
keep the tires operating very close to the slip that yields maximum grip from the
ground. Theoretically, this is equal to limiting static friction, which during braking
tot
Fb,lim = µs (1 − αnlg )(MRW ), (69)
where µs is the limiting static friction coefficient between tires and ground and αnlg
is the fraction of the total aircraft’s weight supported by the nose landing gear under
static conditions. The braking force obtained by evaluating Eq. 69 with µs = 0.8 and
αnlg = 0.06 is considered along with the forces obtained by evaluating Eq. 64 with
For each of the above braking cases, the braking force per braked wheel is evaluated
using Eq. 65 and assuming that all braked wheels contribute equally to the total
The braking torque per braked wheel is computed through Eq. 66 using the
rolling radius of the appropriate tire from Table 14 (data collected from [105]) whose
125
Table 14: Wheel / Tire Selection (data collected from [105])
Rated Nom. Nom. Rim Rolling Three-
load diam. width diam. radius part
(lbf) (in) (in) (in) (in) name
9,650 27 7.75 15 11.8 27x7.75-15
13,700 30 9.5 14 12.65 30x9.5-14
17,200 31 13 12 12.7 31x13.0-12
21,525 36 12 18 15.2 36x12.0-18
24,100 37 14 15 15.25 37x14.0-15
30,100 40 14.5 19 16.65 40x14.5-19
36,800 40 14.5 19 16.65 40x14.5-19
41,100 44.5 16.5 21 18.55 44.5x16.5-21
44,200 46 18 20 18.85 46x18.0-20
51,100 46 18 20 18.85 46x18.0-20
56,600 49 19 22 20.2 49x19.0-22
68,500 54 21 23 22.5 54x21.0-23
72,200 54 21 24 22.2 54x21.0-24
Table 15: Heat sink dimensions vs. tire wheel diameter [97]
Rim Face Rim diam.
diam. width. - OD
(in) (in) (in)
14 2.312 2
15 2.437 2
16 2.5 2.25
17 2.625 2.25
18 2.812 2.25
19 2.875 2.5
20 3 2.5
21 3.125 2.5
22 3.312 2.5
23 3.312 2.625
24 3.5 2.625
25 3.625 2.625
rated load exceeds that computed for the aircraft’s MRW. To find the required axial
force per braked wheel through Eq. 68, the brake dimensions are first evaluated
using the data in Table 15. With the rim diameter selected from Table 14, the
brake outer diameter OD = 2ro is obtained from the third column of Table 15. The
126
inner diameter is then computed from the second column of Table 15 as ID = 2ri =
OD−2 (face width). Evaluation of Eq. 68 with these dimensions yields the maximum
end load F0 that must be applied to actuate the brakes. The required maximum
actuation rate vmax is computed based on the closure of a specified running clearance
The mass of the wheel braking system comprises the mass of the heat-stack Mhs , the
mass of the brake actuation system Mact , and the mass of associated electronics Melec
A significant fraction of the mass of the braking system is comprised of the mass of
the heat-stack Mhs , which is driven directly by the requirement that the necessary
The recent industry trend has been towards the use of carbon brakes as opposed
to steel brakes, since carbon brakes have a higher specific heat capacity chs which
allows the same kinetic energy to be absorbed by a much smaller mass for the same
temperature rise [107]. While it is clear that the use of carbon brakes may offer
significant weight savings [108], these savings are applicable to conventional hydraulic
with variation in braking system architecture, since both conventional and electric
braking systems may make use of either carbon brakes or steel brakes.
conventional hydraulic braking systems for an Airbus A330 baseline [109]. Inspection
of the data provided by Chai and Mason [110] shows that for commercial aircraft,
127
Table 16: Projected mass savings of electric brakes relative to conventional Airbus
A330 hydraulic brakes - reproduced from [109]
Electric Brakes Hydraulic Brakes
Torque plates 20 kg Torque plates 78 kg
Electronics 21 kg Electronics 39 kg
.
EMAs 207 kg Hydraulics 86 kg
(Wiring) 34 kg (Piping) 110 kg
Total 303 kg Total 313 kg
the landing gear group weight may be taken to be approximately 4 % of the aircraft’s
shows that for small, medium, and large commercial aircraft, brakes account for
a roughly constant 16 % of the total landing gear group weight [97]. With these
percentages and the assumption of MTOW = 242,000 kg for the Airbus A3304 , the
predicted weight of all wheel brakes is 1,549 kg.
As seen from Table 16, for the transition to electric brakes, a weight reduction of
162 kg or 10.46 % of the total brake weight was predicted for the torque plates and
electronics and due to the removal of actuating hydraulics. At the same time, there is
the added weight of the electromechanical actuators. Taking these into account, the
net change in mass of the electric wheel braking system relative to the conventional
hydraulic brakes may be written as
F0
∆Mwbs = −κ1 κ2 κ3 (MT OM) + nbrk (72)
(F/M)ema ema
and (F/M)ema is the force-to-mass ratio of the EMA. For the Airbus A330 case
(Table 16), using the listed EMA mass and the end load computed using Eqs. 64-68
for this aircraft yields (F/M)ema = 7.57 kN/kg. Collins [106] lists an EMA mass
4
Online: http://www.airbus.com/aircraftfamilies/passengeraircraft/a330family/
a330-300/specifications/, accessed July 12, 2015.
128
target for an Electrically Actuated Braking System (EABSys) of 18-20 kg for an end
load of 150-160 kN, yielding a comparable (F/M)ema range of 7.50 - 8.89 kN/kg.
For hydraulically actuated brakes, the effective piston area Aef f may be obtained
from the maximum end load F0 and the available working pressure Pwk as
F0
Aef f = . (73)
Pwk
If the brake is to be applied at a rate v, then the required flow rate to the brake is
given by
F0
Qreq = Aef f v = · v. (74)
Pwk
For the case of hydraulically actuated brakes, the maximum flow requirement occurs
during the brake-fill period where the actuator moves at maximum rate v = vmax to
In the case of electrically actuated brakes, the peak power requirement occurs
during ABS operation [109] since combinations of large load and high rate occur. To
obtain the power required from the bus, the corner power may first be obtained as
F0 vmax
Pcp = , (75)
ηpe ηem ηme
where ηpe , ηem , and ηme are the efficiencies of the power electronics, electric motor,
and mechanical drive of the EMA respectively. The corner power is the product
of the maximum load and the maximum rate, which is an operating point that lies
outside an actuator’s operating envelope. Therefore, the peak power required may
The Thrust Reverser Actuation System (TRAS) is responsible for the deployment
and stowing of the thrust reverser mechanism. Typical deployment and stowing times
129
may be around 2-3 seconds. Under nominal conditions, the TRAS is energized only
once per flight and only for these short durations. Therefore, the direct influence of
its power requirements on engine fuel consumption is minimal. However, the power
requirement, though brief, can be significant, and thus may affect the sizing and
TRAS designs fall into two main categories [111]. In the cascade type design,
movable translating sleeves translate aft to uncover cascades through which airflow
(typically secondary airflow) is deflected by blocker doors. In the pivoting type design,
the airflow is deflected through four pivoting doors which also act as the deflectors
for the airflow. The cascade type design has lower actuation loads and is the more
Among the subsystems, the TRAS is unique in that designs over the years have
made use of each of the three secondary power types (pneumatic, hydraulic, and
electric). Early cascade type TRAS designs used pneumatic power [85], where a
pneumatic Power Drive Unit (PDU) provided pressure-regulated bleed air to an air
motor which drove ballscrew actuators synchronized using flexible driveshafts [111].
The majority of TRAS designs use hydraulic power, and feature either separate
linear hydraulic actuators [85] or separate ballscrews which are driven by a common
hydraulic motor (or Power Drive Unit) [112]. More recently, an Electric Thrust
Reverser Actuation System (ETRAS) of the cascade type has been used in the Airbus
A380 [85], in which an electric PDU drives ballscrew actuators through a flex-shaft.
the open literature. A flow rate for the CFM 56 engine is available from work by
Scholz [113], and the electrical power rating (in kVA units) for the Airbus A380
ETRAS is available from [85]. Based on only these two data points, the power
130
requirement per engine for the TRAS is expressed as a linear function of the rated
sea-level thrust of a single engine as
Ptras [kVA or kW] = 0.047 TSL [kN] + 10.782, (per engine) (76)
in which the power is assumed to be in kVA (kilo volt amperes) for electric TRAS and
in equivalent kW for hydraulic TRAS. In the latter case, the equivalent kW figure is
converted into a corresponding flow requirement using the hydraulic power equation.
Since both hydraulic and electric TRAS as considered in this dissertation comprise
ballscrew actuators connected to the movable panels, the control volume for mass
estimation is simplified to include only the hydraulic or electric PDU, whose mass is
estimated as
Ptras
Mtras = , (per engine) (77)
(P/M)pdu
where (P/M)pdu is the power to mass ratio of the hydraulic or electric PDU. Without
this simplifying assumption, the mass of the ballscrew actuators would have to be
taken into account for both the hydraulic and electric TRAS. Since these actuators
are common to both hydraulic and electromechanical TRAS designs, this would not
affect the change in mass due to the electrification of the TRAS.
The ETS concept involves using electric motors to drive the landing gear wheels of the
aircraft in order to permit the aircraft to be taxied without use of the main engines.
The power requirement of the motors is supplied by the APU. The motors may be
integrated either in the nose gear or the main gear. Installation in the nose gear may
be more convenient from an integration perspective, however the tractive force that
can be generated is limited (especially under low grip conditions) by the relatively
low fraction of the total aircraft weight supported by the nose gear (8-15 %).
131
The installation of an ETS always entails a weight penalty due to the incorporation
of the required hardware - electric motors and their power electronics, reduction
gearboxes, and electrical cabling. In addition to this, the APU generator may need
further weight increment. The weight penalty results in additional fuel burn in flight,
when the system is essentially dead-weight. However, the possibility of net fuel savings
arises if the fuel saved through using the system on the ground exceeds the additional
fuel burnt carrying the system in flight. The fuel savings increase with increased
taxiing time, and deteriorate with increasing trip distance. Thus, the system is most
attractive for smaller aircraft which typically fly shorter stage lengths and spend a
on airline demand [32]. Sizing of an ETS and simulation of its power requirements
during taxiing were addressed in greater detail in Chakraborty et al. [114]. Here, only
The ETS operational requirements include the ability to produce the acceleration
necessary to achieve a certain taxiing speed in a specified time, the ability to maintain
a specified taxiing speed, and the ability to break away (commence motion from a
stand-still) against a specified opposing gradient. The weight of the aircraft must also
power electronics.
Airbus from the Airbus Electric Green Taxiing System (EGTS) test program [32]. A
132
Table 17: Electric taxiing system (ETS) requirements
Scenario Initial Final Time Gradient Loadout
speed speed interval
1 Achieve breakaway 0 kt - - 1.5 % MRW
2 Acc. for runway crossing 0 kt 10 kt 20 sec - MTOW
3 Acc. to max taxi speed 0 kt 18 kt 90 sec - MTOW
4 Maintain max taxi speed - 20 kt - - MTOW
the following generalized equation of motion for a taxiing aircraft was proposed in
Chakraborty et al. [114]. It is applicable to aircraft being taxied using the main
engines or an ETS or being towed, and expresses the acceleration V̇ (t) of the aircraft
and the instantaneous mass M(t) during taxiing:
2
T (t) + Ftow (t) − D(V (t)) + (τ (t) − τb (t))
Rr d
− µrr M(t)g − M(t)g sin θ
V̇ (t) = I
M(t) + Rf2 + 2 RIr2
f r
Z t
M(t) = M0 − (F Feng (t) + F Fapu (t)) dt (78)
0
In Eq. 78, F Feng (t) and F Fapu (t) are the fuel flow rates of the main engines and the
APU respectively. The purpose of this equation is to simulate taxiing operation and
certain simplifications may be made for the purpose of system sizing. For example,
for ETS operation, the thrust of the main engines T (t), the towing force Ftow (t), and
the braking torque τb (t) in Eq. 78 may be set to zero. Further, the contribution of
the aerodynamic drag term D(t) is typically negligible at taxiing speeds (≈ 20 kt).
Additionally, the inertia contributions of the aircraft wheels (If /Rf2 and Ir /Rr2 ) are
133
The incremental term ∆µrr (1−V /V ∗ ) models the so-called break-away friction. Thus,
the effective friction coefficient that must be overcome to begin motion is µb/a = µrr,0 +
∆µrr . In Chakraborty et al. [114], the coefficients µrr,0 and ∆µrr were estimated based
on published data for taxiing and break-away thrust settings for a variety of aircraft
taxiing at different percentages of their maximum takeoff weight (MTOW) [115]. The
mean values were found to be µrr,0 = 0.013 and ∆µrr = 0.009, yielding a break-away
followed in order to determine the minimum overall motor power rating that would
dynamics of Eq. 78, the torque-speed envelope of the electric motor was considered,
after accounting for maximum voltage, maximum current, and maximum power
constraints. The minimum required peak power output was found to be related
to the aircraft Maximum Takeoff Mass (MTOM) as
max
Pmech [kW] = (4.10−10)(MT OM [kg])2 + 0.0016 (MT OM [kg]) − 2.2971. (80)
The Airbus Electric Green Taxiing System [32] used two 50 kW motors (i.e., 100
to note that setting this MTOM in Eq. 80 yields a total output power requirement
of 125 kW. The larger power predicted by Eq. 80 is due to the fact that all the
The peak electrical power demand corresponding to the peak mechanical power
demand is obtained by accounting for the efficiencies of the electric motors and the
power electronics:
max
peak Pmech
Pets = (81)
ηem ηpe
134
This peak power requirement affects the sizing of power distribution elements and
also the capacity of the APU generator (which is responsible for supplying the ETS
The main hardware components within the control volume of the ETS are the electric
motors, the power electronics, and the reduction gearboxes. The motor and power
where (P/M)em and (P/M)pe are the power-to-mass ratios of the electric motor and
the power electronics, and ηem and ηpe their respective component efficiencies.
To estimate the mass of the gearbox, the maximum output torque of the gearbox
(required to achieve break-away) is first determined from the total break-away force
where Rroll is the rolling radius of the tire (determined after tire selection, Sec. 4.4)
and Ndrive is the total number of driven wheels. The gearbox mass is computed as
κgb τb/a
Mgb = , (84)
(τ /M)gb
where (τ /M)gb is the torque-to-mass ratio of the reduction gearbox and κgb = 1.5
represents a margin of safety for torque. The total ETS hardware mass is given by
In addition to this mass, there may be a further increase in mass due to the need to
up-size the APU generator. The magnitude of the up-sizing is not determined by the
presence of ETS alone, but also by the solution chosen for the ECS. Sizing of the APU
135
Figure 24: Estimated ETS mass addition as function of Maximum Takeoff Mass
(MTOM). Technology assumptions: (P/M)em = 2 kW/kg, (P/M)pe = 1.05 kW/kg,
(τ /M)gb = 107 Nm/kg, κgb = 1.5, APU generator P/M = 2.76 kW/kg, APU effective
P/M = 1.83 kW/kg
with the relevant technology assumptions. It is interesting to note that with either
assumption (generator re-sizing or APU re-sizing), the predicted total mass addition
for an Airbus A320-sized aircraft (MTOM = 78,000 kg) is around 400 kg, which was
In a full electric taxiing scheme [32], the ETS can be used to push back from the
gate and taxi out to a point near the runway where the main engines are started.
Up to this point, only the APU is active and burns fuel. The main engines must be
run for a certain warm-up time before the aircraft can take the runway, during which
time the APU may be turned off. Similarly, upon arrival, the engines must be run
for a certain cooling-down time before they are turned off. Following this, the APU
supplies power to the ETS to taxi to the gate. From this sequence of operations, the
136
fuel consumed during taxi-out and taxi-in may be readily computed as
where F Fapu and F Fengs are respectively the fuel flow rates of the APU and the main
engines. For the A320 aircraft, these fuel flow rates were stated to be 12.5 kg/min
for both main engines combined and 2 kg/min for the APU [32]. The fuel saved due
from which it is clear that the fuel savings on the ground is directly proportional to
the duration of use of ETS (Tets ), in other words on the total taxiing time.
The fuel penalty due to the ETS mass is computed directly during the mission
performance analysis. However, some insight may also be gained from an analytical
expression estimating this fuel penalty. A manipulation of the Breguet range equation
yields the fuel penalty in flight due to the mass of the ETS for a flight distance R as
(T SF C) R
∆Mf,f light = Mets e V (L/D) −1 , (88)
where T SF C is the thrust-specific fuel consumption of the main engines, V the flight
speed, and L/D the lift-to-drag ratio at a representative cruise condition. Thus, it
is clear that the fuel penalty in-flight increases exponentially as the flight distance R
which justifies the preference to use ETS for aircraft that fly frequent but short flights
137
4.7 Environmental Control System (ECS)
ventilation for the aircraft cabin. It is the single biggest consumer of secondary
power in commercial aircraft. The conventional ECS is pneumatic, and uses bleed
air extracted from the engines for cabin pressurization and thermal regulation. An
electric ECS design is found on the Boeing 787 aircraft [74], where it has facilitated a
predicted a cruise fuel consumption improvement in the range of 1-2 % [116]. The
power requirements of an electric ECS design are considerable, being in the range of
1.14 kW per passenger [117]. The incorporation of Cabin Air Compressors (CACs),
motors to drive them, and the associated power electronics results in significant
mass addition to the aircraft and requires considerable up-sizing of the electrical
PGDS [118]. The development of light-weight power electronics (> 4 kW/kg) has
been identified as one of the major challenges for electric ECS [118]. The tradeoff
between electric and conventional ECS is thus a complex one since it features the
counteracting effects of a reduction in bleed requirement, an increase in shaft-power
The aircraft cabin is typically divided into several zones, each of them having an
independent zone temperature controller. A master cabin temperature is set for the
entire cabin, which can be modified within certain limits for each cabin zone to set
the target temperature for each zone temperature controller. The cabin zone with
the lowest target temperature defines the temperature of conditioned air supplied to
the cabin. The temperature target for the remaining zones is met by blending hotter
trim air with the supply air to obtain the correct temperature. To explicitly account
for the trim air requirements of individual cabin zones, a detailed thermal simulation
138
6500
6000
5000
4500
4000
parameters enter into such an analysis, which are not necessarily available during
throughout the flight for each of the three baselines. A lumped thermal model analysis
is performed by assuming this temperature to be uniform throughout the cabin.
To limit the difference between internal cabin pressure and external ambient
pressure (which generates a radially outward stress on the fuselage), the cabin pressure
As seen in the Boeing 787, the current trend is toward the use of higher minimum
cabin pressures (lower maximum cabin altitudes, reduced from 8,000 ft to 6,000 ft)
to facilitate increased passenger comfort. The pressurization schedule used in this
dissertation is based on that given by Nelson [74] for the Boeing 787 aircraft (Fig. 25).
The nominal volume flow rate of air entering the cabin in modern commercial
aircraft is approximately 20 cfm (cubic feet per minute) per occupant [13]. Thus, the
139
net volume flow rate of air entering the cabin is given by
where Nocc is the number of cabin occupants. This volume flow rate may be converted
to a nominal mass flow rate based on the cabin pressure and temperature as
Pcab V̇nom
ṁnom = , (91)
Rair Tcab
where Rair is the gas constant for air. Since the cabin pressure Pcab reduces with
aircraft altitude, the required mass flow rate for a given volume flow rate requirement
Due to the temperature difference between the cabin temperature Tcab and the
ambient temperature T∞ , heat transfer Q̇loss occurs through the cabin wall (defined
as positive if heat flows from the cabin to the ambient). This may be modeled as heat
flow through three thermal resistors in series: convection between the cabin zone air
at temperature Tcab and the inside wall, conduction to the outer skin of the aircraft,
and convection to the ambient at temperature T∞ . The net thermal resistance Rtot
is then given by
where hint and hext are the convection coefficients for the internal and external
convection processes respectively, and Aint and Aext are the corresponding heat
exchange areas. A value of hint = 5 W.m−2 .K −1 is assumed for still air convection.
representative material properties for paint, insulation, and metallic skin. The heat
140
transfer between the cabin and the ambient through the cabin wall is given by
Tcab − Taw γ−1 2
Q̇loss = , Taw = T∞ 1 + Rc M∞ , Rc = 1 − 0.99 (1 − P r 0.5 ),
Rtot 2
(93)
where Taw is the adiabatic wall temperature corresponding to the freestream Mach
in an internal heat load Q̇int . The metabolic heat load per passenger is assumed to be
same assumption is made for the galley loads, Q̇galley /Npax = 320 W [120]. Taking
into account both Q̇loss and Q̇int , the total ECS heat load is expressed as
Clearly, if the ECS has to remove heat from the cabin, then Q̇ecs < 0.
To maintain the cabin at a desired temperature Tcab against the heat load Q̇ecs ,
air at temperature Tin must be supplied by the ECS to the cabin. The application of
Q̇ecs
Q̇ecs = ṁnom Cp (Tin − Tcab ) =⇒ Tin = Tcab + . (95)
ṁnom Cp
then a higher mass flow rate must be introduced in order to service the heat load. To
lim
determine this mass flow rate, the inlet temperature in such cases is set to Tin = Tin ,
lim max lim min
where Tin = Tin or Tin = Tin depending on which bound was violated. The
Q̇ecs
ṁcab = lim
. (96)
Cp (Tin − Tcab )
141
Thus, it is clear that ṁcab = ṁnom if the solution of Eq. 95 satisfies the temperature
bounds, and ṁcab > ṁnom otherwise.
For modern commercial aircraft using high bypass ratio turbofan engines, the
fuel penalty associated with extracting engine bleed air to supply 100 % of the
cabin airflow requirement would be severe [13]. Therefore, ECS designs employ
recirculation, in which a certain fraction ǫrecirc of the air supplied to the cabin
is actually recirculated air which has previously been extracted from the forward
section of the fuselage by recirculation fans located beneath the cabin floor [13]. The
remaining (1 − ǫrecirc ) fraction of the air is either extracted from the aft section of the
cabin and exhausted overboard, or lost through a slight amount of leakage through
the fuselage. Thus, taking recirculation into account, the required mass flow rate per
between recirculated air of mass flow rate ǫrecirc ṁcab assumed to be at temperature
Tcab , pack discharge air of net mass flow rate (1 − ǫrecirc ) ṁcab at temperature Tpack ,
and cabin supply air of mass flow rate ṁcab and temperature Tin :
5
All analyses in this dissertation assume Npack = 2, which was found to be applicable to most
modern commercial aircraft of the sizes considered based on publicly available information.
142
Figure 26: Component layout within thermodynamic model of ECS pack
In case of conventional (pneumatic) ECS, this mass flow rate is extracted from the
engines as bleed air. Therefore, for Nop,eng operating engines, the bleed air required
For an electric ECS, this mass flow rate is processed by all operating CACs combined.
Since this mass flow rate is admitted as ram air for an electric ECS design, it results
The layout of components within the Environmental Control System (ECS) pack was
143
public domain [121, 122, 123, 124, 125, 126], and is shown in Fig. 26. While electric
ECS solutions may feature customized ECS pack designs [49], conventional bootstrap
units may also be considered within an electric ECS scheme [43]. Therefore, for
simplicity the same pack architecture is assumed for both conventional and electric
The operation of the ECS pack involves the transfer of heat from a supply stream
to a cooling stream. For both conventional and electric ECS architectures, ambient air
admitted through ram air inlets serves as the cooling stream. In case of a conventional
(pneumatic) ECS design, the supply stream (referred to as such since this air is
ultimately supplied to the aircraft cabin) is obtained from engine bleed air. In an
electric (bleedless) ECS design, ambient air admitted through additional dedicated
The supply stream of mass flow rate ṁp enters the ECS pack at pressure P1 , and
temperature T1 . It is cooled in the primary heat exchanger (P-HE) to temperature
T2 , and due to the pressure drop in the P-HE, the pressure drops to P2 . Based on the
setting φ ∈ [0, 1] of the temperature control valve (TCV), the supply stream is split
into two streams: (i) the air cycle machine (ACM) stream of mass flow rate (1−φ)ṁp ,
pressure P2a , and temperature T2b , and (ii) the bypass stream of mass flow rate (φ)ṁp ,
pressure P2b , and temperature T2b . The ACM stream passes into the ACM compressor
(C), where it is compressed to pressure P3 , resulting in a rise of temperature to T3 .
At this elevated temperature, the supply stream rejects heat in the secondary heat
exchanger (S-HE) and is cooled to temperature T4 . The pressure drop through the
S-HE results in the pressure falling to P4 . The supply stream then passes through the
to T5 . At this point, the ACM stream is mixed with the bypass stream, resulting in
pack discharge temperature T6 and pressure P6 . The work extracted from the supply
144
The cooling stream, initially at ambient pressure P∞ and temperature T∞ enters
through a ram air door, with the position of the ram air door actuator (RADA)
determining the admitted mass flow rate ṁr (the RADA is marked as X in Fig. 26).
The cooling stream passes through a diffuser, where its pressure and temperature rise
to P1r and T1r respectively. After accepting heat from the supply stream in the S-HE,
the cooling stream temperature rises to T2r and due to the pressure drop through the
S-HE, its pressure drops to P2r . The cooling stream then accepts further heat from
the supply stream in the P-HE, and is assumed to leave the control volume of the
Associated with each of the pack components, there are a set of well-known
constitute the thermodynamic model of the ECS pack. For solving this model, the
following parameters (color-coded green in Fig. 26) are assumed to be known:
1. The state of the supply stream entering the pack, ṁp , P1 , T1 . The mass flow
rate ṁp is obtained from the cabin airflow requirements and the recirculation
T1 are obtained from the precooler exit conditions. The pressure schedule P1
computed based on the cabin thermal analysis and the recirculation fraction
allowing for a certain percentage pressure drop in the mixing manifold and
145
The pack controller makes use of two controls (color-coded red in Fig. 26) for
adjusting the pack discharge temperature T6 : (i) the ram air mass flow rate ṁr and
(ii) the TCV setting φ. To lower T6 , the RADA opens the ram air inlet more and the
TCV closes more. The reverse occurs if the temperature T6 must be raised [122]. The
RADA also modulates ram air flow to regulate ACM compressor outlet temperature
T3 [126]. However, due to the thermal inertia of the heat exchangers, the effect of the
In general, to minimize the required amount of ram air (and thus ram drag),
the TCV remains almost closed, except when the compressor exit temperature T3
reaches an upper threshold. In such cases, the flow of ram air ṁr is increased to
regulate the compressor exit temperature T3 while the TCV is opened as required to
regulate the pack discharge temperature T6 . Thus, the pack thermodynamic model is
initially solved as a root-finding problem for the ram air mass flow rate ṁr for each
flight condition. For solutions where the compressor exit temperature T3 exceeds a
maximum permissible threshold of T3max = 230◦C [121], the pack model is re-solved
as an optimization problem: minimize ram air mass flow rate ṁr while satisfying the
maximum compressor exit temperature constraint T3max in addition to the pack exit
temperature requirement T6 .
As discussed previously, the supply air mass flow rate ṁp is directly driven by the
cabin airflow requirements. For the conventional (pneumatic) ECS, this mass flow rate
is extracted as bleed air from the engines, which therefore imposes a fuel consumption
146
For the electric ECS, the same mass flow rate ṁp is admitted at pressure P∞ and
temperature T∞ through dedicated ram air inlets and passes through a diffuser. The
where ηd is the efficiency of the diffuser. This air is then compressed by the cabin
air compressors (CAC). Assuming an exit pressure Pcac,out , the corresponding exit
temperature is given by
γ−1
1 γ
Tcac,out = Td,out 1+ Πcac − 1 , (103)
ηcac
where ηcac is the CAC adiabatic efficiency and Πcac = Pcac,out /Pd,out is the CAC
pressure ratio. The total electrical power requirement of the CACs is given by
where ηem and ηpe are the overall efficiencies of the electric motors and power
into the ECS pack, and therefore from Fig. 26, P1 = Pcac,out , T1 = Tcac,out . It is clear
that the electrical power requirement of the CACs depends on the CAC discharge
temperature, which in turn depends on the CAC discharge pressure through Eq. 103.
For a conventional ECS architecture, representative values for the pack entry pressure
and temperature are P1 = 250 kPa, T1 = 180◦ C. However, attempting to match either
this pressure or temperature results in an excessively large CAC power requirement,
especially at cruise conditions due to the high pressure ratio Πcac . Instead, the CAC
discharge pressure is scheduled such that a pressure differential with respect to the
cabin pressure is maintained, which reduces linearly from 130 kPa at sea-level to 100
147
kPa at 45,000 ft cruise altitude. With the ECS model as described, the CAC power
requirement at cruise conditions evaluates to approximately 1.25 kW/occupant. This
compares reasonably well with the approximate figure of 400 kW for a 350 PAX
The operation of the ECS results in the generation of ram drag as external ram air is
admitted into the aircraft. For both conventional and electric ECS solutions, ram air
is used as the cooling stream. In case of electric ECS, the supply air is also external
ram air. The ram drag penalties are estimated pessimistically by assuming complete
momentum loss, in which case the ram drag penalty respectively for the pneumatic
The assumption of complete momentum loss is a conservative one for the case of
the cooling stream ram air, since after passing through the heat exchangers this air
Since the same pack design is assumed for both the pneumatic and electric ECS
solutions, there is no variation in the mass of the ECS packs between the two
of the air conditioning system mass. This includes the mass of the ECS packs for the
conventional architecture, which is therefore used directly. For the electric ECS, the
148
additional mass of the CACs, motors, and the power electronics was computed as
max
Pcac P max /Ncac/pack
Mcac = = elec ,
(P/M)cac (P/M)cac
max 1 1
Mcac,em + Mcac,pe = Pcac + ,
(P/M)em ηem ηpe (P/M)pe
∆Me−ecs = κecs Npack Ncac/pack (Mcac + Mcac,em + Mcac,pe ) , (106)
where (P/M)cac , (P/M)em , and (P/M)pe are the power-to-mass ratios of the CACs,
motors, and power electronics respectively, with ηem and ηpe being the overall
efficiencies of motors and power electronics. The factor κecs accounts for the masses
of other smaller pack components that are not explicitly estimated, and is assumed
to be κecs = 1.25 for all analyses. For component sizing, it is assumed that each ECS
pack can satisfy the requirements of the whole aircraft in case of failure of the other
pack. Each ECS pack is assumed to have two compressors (Ncac/pack = 2) [74], each
Accretion of ice on aircraft surfaces in flight occurs due to the tendency of cloud
as supercooling. This is an unstable state and the supercooled water may readily
undergo a phase change from liquid to solid (i.e. ice) due to agitation by a
passing aircraft. The effects of unmitigated icing on the aircraft may include
loss of aerodynamic efficiency, engine power, adequate control authority, and even
brake and landing gear operation, outside vision, flight instrument functions, and
communication [128] Most commonly, the leading edges of the wings and engine
Protection System (CIPS). In certain aircraft, empennage surfaces may also require
149
ice protection, however this is not universal and is therefore not considered. Ice
protection solutions for the cockpit windshield, drain masts, pitot probes, etc. is also
not considered.
Only thermal ice protection solutions, in which thermal energy is used to prevent
ice accretion, are considered in this dissertation, as they are almost universal on
modern commercial aircraft. In conventional IPS designs, hot bleed air provides the
are found on the Boeing 787 [74, 116]. The IPS may be designed as an anti-icing
system, where the build-up of ice is prevented completely, or a de-icing system, where
ice is allowed to form and then subsequently dispatched [129]. Anti-icing designs
may further be evaporative, where the supplied heat flux is sufficient to completely
evaporate all impinging water and maintain a dry surface or (ii) running-wet, where
the supplied heat flux is only sufficient to melt any ice, but not vaporize the
The analysis assumes a technology level in which the supplied heat flux for all
four IPS variants listed above can be modulated based on flight and icing conditions
The extent of protection (EOP) determines the dimensions of the protected area. It
is defined by the spanwise and chordwise EOP and the spanwise Center of Protected
Area (COPA) [56]. The relevant geometric parameters are defined in Fig. 27.
150
b/2
ΛLE cwing
∆yeop
cslat
ℓeop
cus
eop
yeop,i
ycopa
yeop,o cls
eop
y
For spanwise dimensions: η = b/2
Figure 27: Geometric parameters for definition of spanwise and chordwise EOP
protection. However, as shown in Fig. 28, the EOP varies from aircraft to aircraft.
For example, on the Airbus A320, Airbus A330/A340, and Boeing 767, ice protection
is provided to leading-edge slats located from approximately mid-wing to wingtip. On
the Boeing 737, Boeing 757, Boeing 777, and Boeing 787, ice protection is provided
to leading-edge slats outboard of the engine nacelle, but excluding the most outboard
slat (two, in case of the Boeing 777). On the Airbus A380, ice protection is only
provided to one slat per wing, located just inboard of the outboard engines. In all
of these aircraft, the inboard leading-edge devices spanning from the side-of-body to
the nacelle location are not provided ice protection.
The EOP must be estimated based on the wing profile and geometry, which have
through icing conditions. In general, thinner profiles catch more water as they cause
a greater diversion of the airstream prior to its incidence on the wing leading edge.
Both spanwise and chordwise EOP are affected by the detailed geometry of the wing,
which is not determined (or frozen) until much later in the design process [56].
Liscouet-Hanke [56] cited a study conducted by Airbus that established that the
Mean Aerodynamic Chord (MAC) cmac had the greatest influence on the spanwise
151
Figure 28: Wing leading-edge Extent of Protection (EOP) for some commercial
aircraft. All spans equalized to illustrate relative spanwise EOP.
152
Figure 29: Spanwise EOP as a function of wing Mean Aerodynamic Chord (MAC)
for a number of commercial aircraft
EOP, which decreased with increasing MAC. In that work, a linear relationship
between the spanwise EOP and the MAC had been assumed. However, for this
Using the spanwise EOP of some common commercial aircraft as obtained from
meters. The latter is used for all analyses in this dissertation given the better
agreement with the data points seen in Fig. 29.
153
Respecting the wingtip constraint (ηeop,p ≤ 1.0), the provisional inboard and outboard
spanwise extremities of the protected area are then computed as
∆ηeop
η̃eop,o = min 1.0, ηcopa +
2
η̃eop,i = ηeop,o − ∆ηeop (109)
Since the provisional interval η ∈ [η̃eop,i , η̃eop,o] need not coincide with the spanwise
extremities of the slats themselves, the following logic is applied: a slat is considered to
be provided ice protection if the intersection of the provisional interval η ∈ [η̃eop,i , η̃eop,o]
and the slat’s spanwise interval η ∈ [ηslat,i , ηslat,o ] exceeds more than 40 % of the slat’s
length. For the aircraft sizes considered, this logic results in ice protection requirement
being deemed necessary for the following slats (referring to Fig. 12): (i) slats 3, 4, and
5 on each wing for the SSA (ii) slats 3, 4, and 5 on each wing for the LTA (iii) slat 4
The actual inboard and outboard spanwise extremities of the protected area (ηeop,i
and ηeop,o ) are set to coincide with the identified protected slats. The previously
computed ∆ηeop (from Eq. 107) and ηcopa (from Eq. 108) are then updated as
The total protected leading edge length ℓwips for both wings combined is obtained by
considering the wing’s semispan b/2, leading-edge sweep ΛLE , and dihedral Γ as
∆ηeop b
ℓwips = 2 · . (111)
cos ΛLE cos Γ 2
154
actual wing profile [56], and for this dissertation are expressed as follows:
ceop cslat
cus
eop (η) = · · cwing (η),
cslat us cwing
| {z }
=0.40
ls ceop cslat
ceop (η) = · · cwing (η). (112)
cslat ls cwing
| {z }
=0.22
For a slat with spanwise extent [ηi , ηo ] and slatwise length ℓslat , the protected
where P(η, c̃us , c̃ls ) is the perimeter of the wing airfoil at span station η and between
upper surface and lower surface chordwise locations c̃us and c̃ls respectively. The
summation of the protected areas of all slats determined to require ice protection. It
should be noted that this area has been determined using only information that is
To perform the IPS analysis using only limited information regarding the system
freestream temperature T∞
3. The temperature of the runback water (if any) is assumed to be the same as the
detailed analysis (e.g., [131]) is required which is beyond the present scope
155
Table 18: IPS thermodynamic modeling parameters
Parameter Symbol Comments
Freestream temp., press., Mach T∞ , P∞ , M∞
Reynolds number Rex Re = ρ∞ V∞ x/µ
Prandtl number Pr P r = µCp,a/k0
Nusselt number Nu Nu = 0.0296 Re4/5 P r 1/3
Temp. recovery factor Rc Rc = 1 − 0.99(1 − P r 0.5)
Adiabatic wall temp. Taw Taw = T∞ 1 + Rc γ−12
M∞2
4. The control volume at a given location is assumed to completely enclose both the
heated area and the droplet impingement area. Thus, both the total supplied
IPS heat flux and the total water catch enter this control volume in entirety
captured through a wetness factor [131]. This rivulet effect is not considered
6. The effect of conduction through the metal skin in the flow-wise direction is not
considered due to its relative magnitude in relation to other heat fluxes [131]
Electrothermal IPS solutions permit the modulation of the heat flux supplied to
the protected surface. However, such modulation is typically not possible in the case
of conventional pneumatic IPS, in which the system is either on or off [57]. This
often results in the supply of heat which is in excess of that required to maintain
an acceptable surface temperature for providing ice protection. For pneumatic IPS
solutions, the possibility of using active control to modulate the rate of bleed air
supplied to the protected areas (thus alleviating the associated bleed-induced fuel
penalty) was investigated by Patricelli [132], who concluded that significant reductions
in bleed air mass flow rate were possible in certain conditions. Therefore, flow
156
modulation for pneumatic IPS solutions was considered in this dissertation, under the
assumption that pneumatic IPS designs of the future may employ such modulation.
The estimation of IPS heating requirements subject to the above modeling and
The required heat flux for ice protection is affected by the rate of water impingement
on the protected surface. The total rate of water impingement (dimension: [M][T ]−1 )
onto a surface with projected area Aproj normal to the flow direction at airspeed V∞
is given by
ṁimp = Em (LW C) V∞ Aproj , (114)
where Em ∈ [0, 1] is the water collection (or catch) efficiency and LW C is the Liquid
Water Content of the air (explained further subsequently). The water impingement
expressed as a mass flux per unit protected area (dimension: [M][L]−2 [T ]−1 ) is
ṁimp Aproj
ṁ′′imp = = Em (LW C) V∞ . (115)
Aprot Aprot
In Eq. 115, the water collection efficiency Em is defined as the ratio of the mass of
water impinging on the surface to the total mass of water present in the theoretical
volume swept out by the body:
ṁimp
Em = (116)
ṁtot
It is a function of aircraft speed, droplet size, and the geometry of the surface of
impingement. In general, higher velocities, larger droplet sizes, and thinner surfaces
are available in AIR 1168/4, from which a simplified approximation was deduced by
Krammer and Scholz [133]:
0.613
V∞
Em = 0.00324 , (117)
t
157
where V∞ is in meters per second and thickness t is in meters. This approximation
is strictly true for a median droplet diameter of 20 microns and an altitude of 10,000
ft, but has less than 10 % error for other altitudes from sea level to 20,000 ft [133].
of mean effective droplet diameter and air temperature within the continuous and
intermittent icing envelopes. The LW C for the two icing envelopes is discussed in
The total heat flux q̇tot required to maintain the protected area at a given surface
temperature Ts is made up of contributions from four heat transfer processes6 :
determination of the total required IPS heat flux q̇tot for that flight condition. The
where h∞ is the external convection coefficient and Taw is the adiabatic wall
6
In some analyses, a fifth heat flux component called aerodynamic heating (q̇aero ) is considered.
However, in these analyses, the expression for convective heat flux uses the freestream static
temperature T∞ . If instead the convective heat flux is computed using the adiabatic wall temperature
Taw (as in this dissertation), then it can be shown that the effect of aerodynamic heating is already
accounted for and a separate aerodynamic heat flux term need not be considered.
158
with the characteristic length x taken to be the mean chord of each slat. The
evaporation mass flux and heat flux q̇evap may be expressed as
h∞
ṁevap = min 0.7 (Pv (Ts ) − Rh Pv (T∞ ), ṁimp )
Cp,a P∞
q̇evap = ṁevap Lv , (121)
where the relative humidity of a cloud is generally taken to be 100 % (Rh = 1) [131].
The sensible heat flux q̇sens and kinetic heat flux q̇kin < 0 may be expressed as
The four heat flux components are estimated for a given flight condition to obtain
the total required heat flux using Eq. 118. The heat flux multiplied by the protected
surface area (obtained by applying Eq. 113 to each protected slat) yields the required
rate of heat supply to provide ice protection for a given flight condition.
between -30◦C and 0◦ C. Requirements for the IPS may be derived from the
FAR §25, Appendix C [77] or EASA CS-25, Appendix C [134], both of which contain
very similar information with regard to icing conditions. Both make a distinction
between two icing regimes: Continuous Maximum Icing (CMI) and Intermittent
Maximum Icing (IMI). Both documents characterize these two icing regimes with
159
1. A chart with curves showing the probable maximum (99 %) LW C averaged
over a standard distance (17.4 NM for CMI and 2.6 NM for IMI) for given
3. A chart yielding correction factors for the LW C obtained (Item 1) for averaging
distances other than the standard distances for CMI and IMI
Traditionally (since 1964), CMI conditions (stratiform clouds) have been applied
to the design of airframe ice protection such as WIPS [135, 39]. The aircraft is
time. The assumption is that wings are affected by flight through continuous icing,
but more tolerant to encounters with more severe intermittent icing [39, 136].
application of FAR §25, Appendix C, designers have typically looked up the probable
maximum LW C from it based on an appropriate temperature and a recommended
MVD, which is typically used for computation of ice accretion amounts [135, 137].
To find the WIPS sizing point, the heat flux estimation approach described in
combinations are selected within the aircraft flight envelope and h − T∞ combinations
are set based on the CMI icing envelope [77, 134]. The sizing point is identified as the
max
flight condition yielding the maximum required total heat flux q̇tot . This corresponds
to the maximum bleed air mass flow requirement in case of a pneumatic WIPS and
7
Median Volume Diameter (MVD): A dimension such that 50 % of the mass volume of water
in the cloud sample is contained in drops larger than the MVD, and the remaining 50 % in drops
smaller than the MVD.
160
(a) Pneumatic ice protection system (b) Electrothermal ice protection system
Figure 30: Pneumatic and electrothermal wing ice protection systems (WIPS).
In a pneumatic IPS, hot bleed air extracted from the compressor stages of the engine
is ducted to the surfaces to be protected and then discharged through piccolo tubes
to transfer heat and maintain the desired surface temperature Ts (Fig. 30(a)). The
hot supply air mixes with heated air in the bay, following which heat transfers to the
inner wall through convection, and to the outer protected surface through subsequent
conduction. Explicit modeling of the internal heat transfer process is impossible
without considering the geometry in detail, especially since the internal convection
coefficient has been shown to vary significantly based on the piccolo tube location and
geometry [138]. Since such detailed geometric data may not be available during early
design, a standard approach to modeling the efficiency of the internal heat transfer is
to assume an overall heat transfer efficiency ηp−wips for the pneumatic WIPS [39, 56].
With this assumption, the required bleed air mass flow rate is computed as
q̇tot Awips
prot
ṁp−wips,tot = , (125)
ηp−wips Cp,a (Tsupply − Ts )
where Tsupply is the temperature of the bleed air supplied to the pneumatic WIPS.
In analyses by other researchers, the value of the overall heat transfer efficiency has
been assumed to be 65 % [39, 56]. The bleed air requirement per engine to supply
161
the WIPS for the case of Nop,eng operating engines follows as
ṁp−wips,tot
ṁbleed,wips = (per engine) (126)
Nop,eng
elegant method presented by Krammer and Scholz [133] which is suitable for initial
analysis without a priori knowledge of the detailed geometry is used. The authors
demonstrated that this simplified approach predicted the power requirements of the
Boeing 787 WIPS with reasonable accuracy. In this approach, the ratio of the total
κps ∈ (0, 1). The parting strips are continuously heated with a heat flux q̇ps = q̇tot .
The remaining area fraction (1 − κps ) is covered by cyclically heated surfaces which
are supplied heat for a duration theat within a cycle of duration tcyc , resulting in an
effective activity ratio κcyc = theat /tcyc . The heat flux provided to the cyclically heated
areas is computed as
ρice ℓice
q̇cyc = {cice (Tref − T∞ ) + Lf } ,
theat
where ℓice is the assumed minimum thickness of ice that must be melted during the
heat-on portion of the cycle in order to allow aerodynamic forces to dispatch the ice
buildup above it. Quantities ρice , cice , and Lf are respectively the density, specific
heat, and the latent heat of fusion of ice. Tref = 273 K is a reference temperature.
Considering both parting strips and cyclically heated areas, the required effective
heat flux q̇e−wips and corresponding electrical power requirement Pe−wips are given by
where ηe−wips is the overall efficiency of the electrothermal ice protection system.
Assumed values for the geometric and operational parameters of the electrothermal
IPS that appear in the above relationships are summarized in Table 19.
162
Table 19: Electrothermal IPS design parameters (based on [133])
For a pneumatic IPS design, the subsystem mass includes the mass of the piccolo tubes
in both wings. Instead of computing this mass explicitly, the estimated anti-icing
system mass MA/I from the aircraft sizing program (FLOPS) is used to determine an
effective mass per unit length for the piccolo tubes of both WIPS and CIPS:
MA/I
(M/L)p−ips = (M/L)p−wips = (M/L)p−cips = , (128)
ℓwips + ℓcips
where ℓwips (from Eq. 111) and ℓcips are the total protected lengths for the WIPS and
CIPS respectively. The mass of the pneumatic WIPS is then computed as
For an electrothermal IPS design, data provided by Al-Khalil [139] which consisted
of the densities and thicknesses of the materials forming the heated area is used to
compute the effective mass per unit area of the heating pads (M/A)e−ips. Using this,
163
4.8.8 Drag Penalty Estimation
A drag penalty is assessed only for the case of electrothermal de-icing, for which it is
where ticing is the icing time (in minutes, = theat /κcyc ), V∞ is the freestream velocity
(in miles/h), c the chord (in inches), T0,∞ is the air total temperature (in Fahrenheit),
and α is the angle of attack. For the maximum local impingement efficiency, β̄m = 0.65
is set based on inspection of the data presented in [140]. The above correlation is
valid both for rime ice and glaze ice. The sectional drag increment computed through
where q̄ is the dynamic pressure, Sw is the wing planform area, and Splan,prot is the
The approach described above yields a scalable estimate for the power requirement
and mass of the WIPS based only on quantities that are available during conceptual
design. A similar approach is used for the Cowl Ice Protection System (CIPS), as
described in the following section.
164
De-icing systems are not considered in this dissertation for the CIPS due to the
risk of dislodged ice being ingested into the engines. Similar to the WIPS analysis,
the CIPS analysis assumes a technology level in which the supplied heat flux for all
four CIPS variants can be modulated based on flight and icing conditions in order to
maintain a desired surface temperature. The sizing and analysis approach employed
for the CIPS is very similar to that described in the previous sections for WIPS.
Therefore, the following descriptions have been abbreviated, with references provided
to the relevant WIPS sections where necessary.
The total protected length ℓcips and total protected area Acips
prot are determined as
Acips
prot = Neng (π Dnac,f ace ) (π ∆xcips ) (133)
where Dnac,f ace is the diameter of the front face of the engine nacelle and ∆xcips is the
axial extent of protection measured along the longitudinal axis of the nacelle. The
protected area is computed assuming a simplified nacelle geometry as shown in Fig. 31,
∆xcips
1
2
Dnac,f ace
165
4.9.2 Estimation of Required Heat Flux
The required CIPS heat flux is estimated for a given flight and atmospheric condition
in the same manner as for WIPS in § 4.8.4, and is therefore not repeated.
Traditionally, IMI conditions (cumuliform clouds) have been applied to the design
of engine (cowl) ice protection systems [135, 39]. Here, the aircraft is considered
The assumption in this case is that ice accretions of even short durations cannot
be tolerated for engine inlets as they may overwhelm the CIPS and cause engine
operating problems [39, 137]. Normal CIPS design practice is to design for at least
a running-wet surface for intermittent maximum conditions if the ice formed during
such an encounter compromises safe flight [137]. The approach taken to find the
CIPS sizing point is similar to that for finding the WIPS sizing point. The heat flux
envelope and h − T∞ combinations are set based on the IMI icing envelope [77, 134]
(as opposed to the CMI envelope used for WIPS). The sizing point is identified
as the flight condition yielding the maximum required total heat flux. Providing
ice protection at this point requires the maximum bleed air mass flow in case of a
pneumatic CIPS and the maximum electrical power in case of electrothermal CIPS.
The CIPS power requirements are estimated in the same manner as the WIPS power
166
4.9.5 Mass Estimation
where (M/L)p−ips is the mass per unit length of the piccolo tubes (Eq. 128). In a
where the mass per unit area (M/A)e−ips was discussed previously in § 4.8.7.
No drag penalties are assessed for any of the four CIPS designs considered, since none
The modeling approaches for the power consuming subsystems (FCAS, LGAS, NWSS,
WBS, ETS, TRAS, ECS, WIPS, and CIPS) were described in this chapter. For each of
them, the control volume for analysis was first defined, following which approaches to
estimate subsystem mass, secondary power requirements, and drag increments (if any)
were presented for each technical solution considered. These quantities are ultimately
fed back into the aircraft sizing and mission performance analysis to obtain the impact
directly drive the sizing of the elements of the power generation and distribution
167
CHAPTER V
DISTRIBUTION SUBSYSTEMS
The power generation and distribution subsystems (PGDS) are responsible for
generating the different types of secondary power and distributing it to the power
consuming subsystems (described in Chapter 4). These include the following:
power extracted from the engine and the distribution of hydraulic power to
power extracted from the engine and the distribution of electric power in the
for the extraction of pneumatic power in the form of pressurized bleed air
from the engine, the regulation of the required air as per requirements, and
its distribution to pneumatic power consuming subsystems
for the extraction of mechanical power (shaft-power) from the engine and the
168
This chapter describes the modeling approach for each PGDS listed above.
As discussed previously in Chapter 3, the determination of PGDS characteristics
Therefore, the first part of this chapter (§5.1) describes how the component
physical constraints. Examples of the latter include the need to deliver a required
flow rate to a hydraulic consumer without exceeding a maximum permissible pressure
without exceeding a maximum permissible voltage drop. The capacities of the power
sources are not obtained by simple summation of the peak power requirements of the
connected consumers. They are instead sized by identifying constraining load cases
where peak consumer demands or the unavailability of similar power sources (through
engine or component failures) maximize the power demand. Further, even though the
PGDS do not directly consume power like the power consuming subsystems, there
increase in secondary power consumption at the aircraft level. The power dissipation
systemic inefficiencies caused by friction losses in piping and ducting or ohmic losses
in electrical cables. The second part of this chapter addresses the modeling approach
for the HPGDS (§5.2), EPGDS (§5.3), PPGDS (§5.4), and MPGDS (§5.5).
169
5.1 Heuristic Determination of Connectivity Among
Subsystem Architecture Elements
connectivity among the elements of the architecture, such as those between prime
movers and power sources, power sources and power systems, and power systems and
power consumers.
Certain similarities were assumed in the associations among hydraulic and electric
power systems and the various actuation functions. Therefore, the definition of
electric system architecture (§5.1.2), the hydraulic system architecture (5.1.3), and
The actuation functions for the FCAS, LGAS, NWSS, WBS, and TRAS can use both
hydraulic (conventional) and electric power (More Electric). For these subsystems,
the layout of actuators and the power systems supplying the actuators is driven by
redundancy requirements and the criticality of the actuation functions. In general,
actuation functions that are highly critical (e.g., actuation of the primary flight control
surfaces) are provided the highest degree of redundancy in the form of multiple
provided by associating multiple power sources with each power system. These
performed much later in the design process. At the same time the redundancy
significantly to the mass of the vehicle. The solution adopted in this dissertation for
170
determining the associations of power systems to the actuation subsystems involves
the identification and subsequent application of a set of heuristic rules whose validity
considered feasible,
investigations into the reliability of electric actuators [141, 29, 31] support the
possibility that these conditions may be realizable. Subject to the central assumption,
powered by n electric power systems. From this in turn, it follows that heuristic rules
well. A brief summary of the heuristic rules identified by inspection of such actuation
1. Ailerons and elevators (FCAS): Control surfaces such as ailerons and elevators,
which are flight-critical, are provided with two actuators per panel. This is also
true if multiple aileron panels exist per wing, and if multiple elevator panels
1
Information regarding fully conventional hydraulic actuation in existing aircraft was determined
from a number of public domain documents available from the website www.smartcockpit.com. This
resource contains freely accessible PDF documents describing various systems of different aircraft
from different manufacturers. For the purpose of identifying heuristic rules, the most relevant were
documents dealing with ATA 27: Flight Controls and ATA 29: Hydraulic Power. Information
regarding MEA architectures was obtained in particular from Refs. [116, 74, 24].
171
exist per horizontal stabilizer. Each actuator is supplied by a single power
system. Let the terms aileron group and elevator group refer to all ailerons
a. If the aileron group and/or the elevator group is powered by the same type of
power (i.e., either hydraulic or electric but not both), then three such power
b. If the aileron group and/or the elevator group is powered by both types of
power (i.e., both hydraulic and electric), then two power systems of each
actuators. If the rudder is partitioned into two panels, then each panel is
provided with three actuators. Each actuator is supplied by a single power
system. Let the term rudder group refer to all rudder panels taken collectively.
(i.e., either hydraulic or electric), then there must be three such power
(i.e, both hydraulic and electric), then it may be supplied by two hydraulic
c. If two rudder panels are present and the rudder group is powered by a single
type of power (i.e., either hydraulic or electric), then collectively the rudder
d. If two rudder panels are present and the rudder group is powered by both
types of power (i.e. both hydraulic and electric), then two power systems
of each type are required. Each panel must be supplied by one hydraulic
172
system and one electric system. The same hydraulic system cannot supply
both panels. The same electric system cannot supply both panels
3. Spoilers (FCAS): Since the presence of multiple spoiler panels provides surface
redundancy, each spoiler panel is driven by only one actuator. Each actuator
panels must always preserve left wing / right wing symmetry, i.e., equivalent
spoilers on both wings must be powered by the same power system. Let the
term spoiler group collectively refer to all spoilers.
a. If the spoiler group is supplied by one type of power (i.e. either hydraulic
or electric), then two such power systems are required (if not initialized
already). These may be associated with the spoiler panels respecting the
b. If the spoiler group is supplied by one type of power (i.e. either hydraulic
or electric) and three systems supplying that type of power are already
initialized, then these may be associated with the spoiler panels respecting
c. If the spoiler group is supplied by both types of power (i.e., both hydraulic
and electric), then two power systems of each type are required. These
a. If the THSA is supplied by only one type of power (i.e., either hydraulic or
173
electric), then two such power systems must supply the THSA
b. If the THSA is supplied by both types of power (i.e., both hydraulic and
electric), then two such hydraulic systems and one electric system must
supplied by two power systems. These may be two hydraulic systems, two
of each landing gear leg is served by a single retraction actuator. The actuator
is supplied by a single power system. In case of three landing gear legs (one
nose gear and two main gears), the same power system supplies the retraction
actuator for each leg. In case body-mounted landing gears are also present,
a separate power system supplies the retraction actuator for these additional
landing gears.
the landing gear actuators for the nose and wing-mounted landing gears.
body-mounted main gear are supplied by two power systems (one primary
system and one alternate system). If body-mounted gears are present, then
the primary system for the wing-mounted gears is the alternate system for the
body-mounted gears. The alternate system for the wing-mounted gears is the
174
10. Thrust reverser actuation (TRAS): Each thrust reverser is powered by a single
power system. If by this point there are less than or equal to two power systems
of the type supplying the TRAS, then (i) for a total of two thrust reversers, a
single power system may supply both reversers, (ii) for a total of four thrust
reversers, a total of two power systems may supply the reversers. Left wing /
right wing symmetry must be maintained in this case. However, if by this point
there are already three power systems of the type supplying the TRAS, then
(i) for a total of two thrust reversers, each may be powered by a different power
system, (ii) for a total of four thrust reversers, the two outboard reversers may
be powered by the same power system, while the two inboard ones are each
The application of the above heuristics in the order in which they are stated allows
2. the association of each end actuation load (an individual actuator) with either
The associations established using these heuristics are used subsequently for the sizing
of the power distribution elements (electric cables and/or hydraulic pipes) during the
Before the elements of the EPGDS can be sized, its architecture must be
defined. Unlike the HPGDS and the PPGDS, which follow reasonably well-defined
architecture possibilities exist for the EPGDS. In particular, several options exist
175
with regard to the choice of voltage type used by major electrified loads (AC vs.
DC) and also voltage magnitude. For MEA, due to the significantly higher electric
load, higher voltage levels are often considered in order to reduce the magnitude of
transmitted currents and thus the cable weight. Thus, while voltage levels of 115 VAC
and ± 135 VDC were common in prior aircraft [142], higher voltage levels such as
230 VAC and ± 270 VDC have already appeared in service [116]. There is also a recent
trend towards the use of variable frequency power generation, in which a generator
is connected directly to the engine. It therefore generates AC power at a variable
frequency that can range between 360-800 Hz [116]. This allows the elimination of a
heavy and complex component called a constant-speed drive, which has traditionally
been used for constant frequency power generation [142]. Further, recent aircraft
off a running engine, or in motor mode using another electric power source to start
the engines, thus eliminating the traditional air turbine starter [116]. The electrical
used as the basis of the EPGDS for the MEA architectures. Some of its important
230 VAC. Each such starter-generator is connected to its own variable frequency
2. 230 VAC power is directly supplied to large loads such as WIPS and CIPS (if
electrified) that are insensitive to the frequency of supplied power [74]. This
also helps to reduce the magnitude of converted power, and thereby the mass
176
Figure 32: Electric system template - the basis for EPGDS architecture
3. 230 VAC power is converted to ± 270 VDC power using AC-to-DC power
in order to supply large loads such as the ECS, electrified actuation functions,
The DC power systems DC1 , . . . DCk shown in Fig. 32 are in fact the electric power
systems described in §5.1.1 for the actuation functions. The association of the DC
power systems to the electrified actuation functions has thus already been described.
being used for each pack. The ETS, if installed, is assumed to be powered only by a
single DC system. This is a valid assumption, since the ETS is neither flight-critical
nor ground-critical, and in the event of its failure, there is always the provision to
taxi conventionally using main engine thrust.
Due to the assumptions stated above, the number of AC buses equals the number
is actually determined taking into account several factors other than merely the
between generator capacity and weight [56] or the tendency of the nacelle diameter
to balloon undesirably if a single high-rating generator is used [42]. In the approach
presented, the number of EDGs per engine Nedg/eng is set according to the following
177
simple rules based on the number of engines Neng and the subsystems electrified:
2. If both WIPS and CIPS are electrified and Neng = 2, then Nedg/eng = 2
3. If primary flight control surfaces are electrified and Neng = 2, then Nedg/eng = 2
The above rules take into account both the relative magnitudes of electrical loads
arising from electrification of different subsystems and the number of engines in order
to determine Nedg/eng . All EDGs are assumed to be identical in terms of rated power
generation capacity.
The heuristic logic for determining the connectivity of the hydraulic power systems
with the hydraulic loads has already been described in §5.1.1. The same logic
also allows the determination of the number of hydraulic systems required for any
actuation architecture, ranging from fully conventional (3) to fully electric (0).
Heuristic rules for the association of hydraulic power sources (pumps) with the
hydraulic power systems were also derived based on the inspection of existing aircraft
hydraulic systems. In general each hydraulic system is supplied by at least two pumps.
Typically, one of these is an Engine Driven Pump (EDP) while the other may in some
cases be an Electric Motor (Driven) Pump (EMP). In such cases, the capacity of the
EMP is considerably less than that of the EDP. In more recent designs however, the
using two EMPs and no EDP (e.g., Boeing 787 Center System [74]). If hydraulics are
present, then there may be one, two, or three hydraulic systems. Since the aircraft
considered have either two or four engines, a total of six combinatorial possibilities
178
Table 20: Association of hydraulic power sources with hydraulic power systems
(EDP: Engine Driven Pump. EMP: Electric Motor Pump. RAT: Ram Air Turbine.
EDP-E1 means EDP driven off Engine # 1. EMPs are simply numbered sequentially.)
arise, for which the associations of hydraulic power sources to hydraulic systems is
summarized in Table 20.
The Ram Air Turbine (RAT) is able to pressurize one of the hydraulic systems in
the event of failure of all engines. Inspection of existing hydraulic system architectures
of Airbus and Boeing aircraft revealed that the hydraulic system supplied by the RAT
1. always supplies one actuator on the rudder, aileron, and elevator panels for
aircraft that have only one rudder, two ailerons, and two elevators (to permit
2. typically supplies one actuator on each outboard aileron for aircraft with
multiple aileron panels per wing. In certain cases, one actuator on the inboard
aileron panels may also be supplied (this latter logic is used for this work, and
3. is always one of the two systems that supply the leading-edge and trailing-edge
4. is typically (but not always) one of the two systems that supply the THSA (for
179
5. is the system that powers gear actuation, wheel braking, and steering functions
(to permit these functions following failure of all engines)
6. does not supply thrust reversers in the case of twin-engined aircraft (presumably
These heuristics and the ones identified previously allow for the heuristic definition
to CIPS
FACV
CIPV
CV
IP PHX
PRSOV
EBV
HP
HPSOV
to WIPS
& ECS
As shown in Fig. 33, bleed air (often referred to by engine manufacturers as customer
bleed ) is extracted from a turbofan engine through bleed ports. In most cases, the
bleed can be tapped from at least two ports. Under high throttle settings that are
commonly used for takeoff, climb, and cruise, an Intermediate Pressure port (IP)
are insufficient, and air is instead extracted from a High Pressure port (HP). In
this case, the High Pressure Shutoff Valve (HPSOV) opens and a Check Valve (CV)
prevents backflow of the HP air into the IP. The port selection logic is an attempt
180
at localized energy management as it attempts to extract air from the least wasteful
bleed port. Even so, under most operating conditions, the temperature and pressure of
the extracted bleed air are too high for direct use by downstream consumers. Passing
such high-temperature air through the wing ducts presents a problem since these
are often in close proximity to fuel lines, and the bleed air temperature it typically
above the auto-ignition temperature of fuel. Excessively high pressure can also cause
sealing and containment problems. Therefore, the temperature of the extracted bleed
air is reduced by passing it through a Precooler Heat Exchanger (PHX), which is an
air-to-air heat exchanger in which the bleed air is cooled against a stream of cooler
air extracted from the fan. The flow of fan air is modulated by a Fan Air Control
Valve (FACV) based on temperature feedback from the bleed air exiting the PHX in
order to attain a target temperature between 180-200◦C. The pressure of the bleed air
is down-regulated to around 300 kPa [118] by a Pressure Regulating Shut Off Valve
(PRSOV), which can also be completely closed to prevent any bleed extraction.
The down-regulated air, which is now at acceptable temperature and pressure can
be supplied to the pneumatic users. If the CIPS is activated, then the bleed air can
pass through the Cowl Ice Protection Valve (CIPV) from where it is ducted to the
nacelle front face. A review of the pneumatic system designs of existing commercial
aircraft showed that in certain cases, bleed air for nacelle ice protection (CIPS) is
tapped prior to the PHX and therefore is hotter than the air supplied to other
pneumatic users such as the WIPS and the ECS. However, in certain other cases, the
air used for the CIPS also passes through the PHX. For all results presented in this
dissertation, the latter design is assumed. A logic switch exists in the PPGDS analysis
module which can be set to enforce the first design, in which case electrification of
the ECS and WIPS results in elimination of the PHX of each engine.
In order to supply the ECS and WIPS, the bleed air passes through the Engine
Bleed Valve (EBV) and is ducted through wing ducts into a pneumatic manifold in the
181
Engine 2 Engine 3
EBV EBV
CBV
Engine 1 Engine 4
EBV EBV
(from APU)
fuselage area (Fig. 34). The pneumatic system is divided into left and right halves,
which are interconnected through a Cross-Bleed Valve (CBV) so that one system can
supply users on the other side in case of a failure in the other system. Under nominal
operating conditions, the left system supplies the left WIPS and the left ECS pack,
while the right system supplies the right WIPS and right ECS pack. Bleed air flow
to the WIPS is controlled by the Wing Ice Protection Valves (WIPV), while flow into
the ECS packs is controlled by Pack Control Valves (PCVs). In addition, since the
APU must be able to support ECS operation on the ground, an APU supply duct
runs from the aft fuselage to the central pneumatic manifold, through which the APU
can supply air for ECS operation on the ground. The APU supply duct is also used to
supply pressurized air that is used by the conventional pneumatic starters to start the
main engines. Engine starting has not been considered in this dissertation. However,
in previous studies which considered ECS electrification, the increased power rating
of the main engine generators was found to be sufficient for these to be used as motors
to start the main engines [42, 43]. Based on this finding, the electrification of the
ECS is deemed to be a sufficient condition for elimination of the APU supply duct,
as described subsequently.
182
Of the subsystems considered, the ECS, WIPS, and CIPS are pneumatic
power users in a conventional architecture. Within the MEA architecture space,
electrification may affect none, some, or all of these subsystems. This results in
pneumatic system and the sizing of its components. The logic rules that are applied
• Case 1 - ECS, WIPS, and CIPS are pneumatic: The layout of the PPGDS is as
described above and shown in Fig. 33 and Fig. 34
• Case 2 - Only ECS is electrified: The APU supply duct is eliminated. Remaining
ducts and precoolers are down-sized based on the reduction in the mass flow
• Case 3 - Only CIPS is electrified: Nacelle ducts running from bleed port to
• Case 4 - Only WIPS is electrified: Wing ducts and central ducts are down-sized
• Case 5 - ECS, WIPS, and CIPS are all electrified: Entire PPGDS is removed
• Cases 6, 7, 8 - Any two among ECS, WIPS, and CIPS are electrified: The
The sizing approach for the pneumatic ducts and the PHX based on the mass flow
183
5.1.5 Translation of Logical Connections to Physical Equivalents
The architecting heuristics described in the preceding sections allow the logical
do not provide any information regarding either the locations of the power sources,
power systems, and power consumers, or the physical paths taken by the power
The locations of the power consumers, systems, and sources are fairly
well-established for both conventional and MEA architectures, and thus may be
specified with reference to the geometry of the parent body that they are contained
to the control surfaces, which in turn are defined parametrically on the lifting
surface (using spanwise and chordwise coordinates)
2. Components which are located on the outer periphery of the engine nacelles (or
may be approximated as being so) are conveniently defined using their clockface
position and distance aft of the nacelle front face (normalized by nacelle length).
This allows the positions of engine-driven pumps and generators, the CIPS
piccolo tube or heating elements, the thrust reverser actuators, and the exit
brakes, and electric taxi motors are defined by their normalized position on the
landing gear legs, which themselves are defined by their attachment points to
The ultimate goal in creating such definitions is to be able to compute the Cartesian
coordinates of each of these components, since they define the extremities (start
184
and end points) of the power distribution elements connecting them. However,
simply knowing the extremities of the distribution elements is not sufficient, since
start and end points. Even in such cases, through inspection of subsystem layouts
of existing aircraft, it is possible to develop and program rules for the paths taken
by the distribution elements to get to the so-called “four corners” of the aircraft.
Such rules allow the parametric definition of one or more intermediate points in
between the starting and ending points which collectively determine the route taken
by the distribution element. For each PGDS, such rules may be defined for paths
connecting an assumed point of origin to all other relevant points. For example, for a
the center fuselage area, parametric routing rules are defined for distribution elements
starting from the origin and ending at the locations of each hydraulic load. Additional
rules are defined for paths leading from the power sources (pumps) to the system
origin. Since no more than three hydraulic systems are envisioned, a set of rules
for each such system may be developed and saved, thus creating a template for that
hydraulic system. If the architecting algorithm (§5.1.1) determines that only two
hydraulic systems are sufficient, then two of these three templates are activated.
Since no one hydraulic system supplies all hydraulic loads, only the system-to-load
connections determined through logical architecting to exist need be developed. This
its equivalent physical connection within the aircraft, as shown in Fig. 35.
Within this geometric model, each power distribution element SA→B connecting
start point A to end point B is represented by a set of n points in Cartesian space, the
first and nth points by definition coinciding with A and B respectively. The curvilinear
185
Figure 35: Translation of logical subsystem connections to physical/geometric
equivalents - VLA-00000 (fully conventional subsystems architecture) is shown. Red
lines indicate the PPGDS. Shades of green represent the three hydraulic systems of
the HPGDS
The masses of the power distribution elements are obtained using the computed
lengths and the elements’ mass per unit length (discussed in subsequent sections).
The power generation elements of the HPGDS are the hydraulic pumps, which suck
fluid from a hydraulic reservoir through suction lines and pressurize it to a high
186
(pressurized fluid flow). Systems with nominal pressures of 3,000 psi were the norm
prior to the Boeing 787 and Airbus A380 aircraft, which have 5,000 psi hydraulic
systems. Due to their successful use in these aircraft and beneficial impact on weight,
it is foreseeable that future hydraulic systems would employ such higher pressures.
Therefore, only 5,000 psi hydraulic systems were considered in this dissertation.
The pressurized fluid is then conveyed through the distribution elements, which
are the hydraulic lines/pipes, to the hydraulic power consumers. In the conventional
architecture (as addressed in this dissertation), these power consumers are the
FCAS, LGAS, WBS, NWSS, and TRAS. Additional components of the HPGDS
include accumulators, numerous regulating and check valves, and associated fittings,
couplings, and brackets. The hydraulic pumps and piping (including associated
fittings and valves) contribute to the mass of the HPGDS. The subsystem dissipates
power due to component inefficiencies and leakage flow. The estimation of HPGDS
mass and power dissipation are discussed in the following sections.
Since hydraulic systems have been in use for several decades, regression equations
for their overall mass exist. For instance, the FLOPS-generated mass breakdown
includes estimated mass of hydraulics, and takes into account the system pressure.
Banel-Caule [144] presented the weight breakdown shown in Table 21 for the weights
187
However, the net hydraulic system mass or even the relative weight percentages
of Table 21 are not sufficient for estimating the mass of hydraulic piping in all MEA
architectures. This is due to the fact that in many of these MEA architectures,
only some of the actuation functions are electrified. Among the actuation packages
considered (Table 3), in packages other than Packages 0, 6, and 7, there is a partial
removal of the conventional hydraulic system. The mass of the remaining (residual)
a component build-up approach. In particular, the masses of the hydraulic pipes are
proportional to their lengths, which are not known with certainty in the early design
phases. In fact, the final routing of these pipes may be quite complex and can often
be affected by the need to avoid other equipment [144]. However Airbus, based on a
study of data from Airbus aircraft [144], established the following in connection with
hydraulic piping:
hydraulic system to each and every hydraulic consumer. Instead, there are
common pipes which go from the generation point to a defined meeting point.
There, they split (branch) to form dedicated lines which then run to the location
2. For a given material and nominal system pressure, the weight per unit length
(W/L) of a filled hydraulic pipe is proportional to the filled pipe’s cross-sectional
area Acs , flow velocity v, and flow rate Q as Q = Acs v, this implies that for a
188
given pipe, the cross-sectional area is proportional to the required flow through
the pipe. This yields Acs ∝ Q =⇒ (W/L) ∝ Q =⇒ W ∝ Q L
4. The weight of hydraulic piping is proportional to the product of the flow rate, the
pipe length, and the number of actuators - summed over all hydraulic consumers.
This was verified by calibrating against real data for members of the Airbus
Performing the summation of Item (4) over all of ‘j’ hydraulic connections rather
than over all hydraulic consumers (a consumer may be served by multiple connections,
X X
Wpiping ∝ Qj Lj =⇒ Wpiping = Kpws Qj Lj . (137)
j j
where the parameter α is the ratio of piping mass to the total hydraulic system mass.
From the percentage weight breakdown shown in Table 21, this parameter is set as α =
0.45 + 0.24 + 0.12 = 0.81. The underlying assumption is that the weights of couplings,
brackets, manifolds, filters, and valves associated with piping varies proportional to
the weight of the piping itself. For each hydraulic consumer considered, the maximum
189
in Chapter 4) and using the hydraulic power equation. Regarding the lengths Lj in
the above formulation, a question arises regarding how the common piping segments
prior to the branching out to different consumers factor into the summation, if at all.
distribution point. It then splits into two branches that run lengths ℓ1 and ℓ2 to two
to the two consumers, including the common length. The continuity equation yields
= Kpws {(Q1 + Q2 ) ℓ0 + Q1 ℓ1 + Q2 ℓ2 }
Equation 139 establishes that even though dedicated (exclusive) piping runs do not
exist from the source to each hydraulic consumer, for the purpose of the summation
of Eq. 138, it is sufficient to consider the direct (non-exclusive) length from the source
to each hydraulic consumer. This result presents a significant advantage since these
direct lengths are readily computed from the 3-D geometric model of the aircraft
(using Eq. 136) following the architecting of the HPGDS (§ 5.1.1 and § 5.1.3).
The piping weight sensitivity factors Kpws for the SSA, LTA, and VLA
architecture for each. For each aircraft, the computed Kpws is assumed to be an
invariant characteristic. Subject to that assumption, the residual piping weight of all
190
MEA architectures for that aircraft is evaluated as
!
X
M EA
Wpiping = Kpws Qj Lj . (140)
j M EA
This approach is superior to one that makes a simple assumption that the residual
piping weight is proportional to either the residual piping length or the residual flow
established previously that the weights of two pipes of equal length carrying different
flow rates or that of two pipes of unequal length carrying identical flow rates will not
The estimation of the weight of the hydraulic pumps (EDPs and EMPs)
necessitates the estimation of their required flow capacities. These capacities are
determined using a hydraulic load analysis [113] that includes both nominal and
failure scenarios. In the scope of this dissertation, only a limited number of load
cases are considered, which are summarized in Table 22. Each column represents a
different load case, for which the load demands of each of the conventional hydraulic
Table 22: Hydraulic load cases considered for pump sizing (Abbreviations - G:
ground, TO: takeoff, F-U: flaps up/retract, LG-U: landing gear up/retract, CR: cruise,
DE: descent, F-D: flaps down/extend, LG-D: landing gear down/extend, S/B: spoilers
extend + braking, B/TR: braking + thrust reverser deployment)
191
consumers is specified as a fraction of its maximum flow (therefore, 1.0 implies the
consumer’s maximum flow demand). Since EDPs are geared to the accessory gearbox,
the available flow varies with the engine N2 speed. Considering i = 1 . . . n hydraulic
where ǫ is the ratio of the flow capacity of the EMP to that of the EDP for hydraulic
systems that are pressurized by one pump of each type. The flow availability of the
EMP does not vary with engine speed. For systems pressurized by two EDPs, the
factor ǫ may simply be omitted in Eq. 141. For systems pressurized by two EMPs,
the denominator can neglected entirely.
Though basic, this pump sizing approach can account for the fact that the
constraining load case may change depending on which actuation functions are
dependence of the available pump flow on the engine speed, the sizing load case
is typically one where high flow-demand loads such as landing gear, flaps, or thrust
reversers must be actuated with the engine at or close to flight idle [17]. However,
if these actuation functions are electrified, then the constraining load case for pump
sizing may change. Such changes are captured in the approach described above.
Once the required pump capacity is determined, the masses of the pumps are
determined based on power-to-mass ratios that were identified from various pump
0.70 - 0.76 kW/kg. Since hydraulics have reached technology saturation, the upper
values from these ranges are taken to represent the SOTA in pump technology, and
192
5.2.2 Power Dissipation
where the first term represents the power dissipation in the hydraulic pumps with
overall efficiency ηpump < 1, and the second term represents the power dissipation due
to the need to maintain nominal system pressurization ∆pnom against a leakage flow
Qlkg . For simplicity, the leakage flow is modeled as a fixed fraction of the installed
hydraulic capacity based on the hydraulic load profile for the Airbus A320 [145]. The
power dissipated within the HPGDS manifests itself at the architecture level as an
already provided by aircraft sizing tools (such as FLOPS). Since the electrification of
subsystem architectures necessarily involves the expansion of the EPGDS, the mass
of the EPGDS for a MEA architecture is computed by adding the subsystem’s mass
With reference to the EPGDS template shown in Fig. 32, these major components
are seen to be (i) Engine Driven Generators (EDGs) and APU Generators (APUG),
which generate electrical power, (ii) ATRUs (power conversion equipment), which
transform 230 VAC power to ± 270 VDC power, and (iii) electrical feeders and cables
which transmit electrical power between EPGDS components and loads. Thus, the
in which M0,EP GDS is the estimated mass of the EPGDS for the conventional
193
subsystem architecture. The following sections describe the estimation of the mass
increments ∆Mgen , ∆Matru , and ∆Mcbl due to increased generator capacity, additional
max gnd
Papug = Papug,0 + Pets + Peecs ,
where Papug,0 is the APU generator capacity of the conventional baseline architecture,
max gnd
Pets is the maximum power demand for electric taxiing (if equipped), and Pecs is
the ground power requirement for electric ECS. If there is no ETS installed and the
ECS is not electric, then the APU generator capacity remains at the baseline value.
The required capacity of the Engine Driven Generators (EDGs) is computed using
1. Due to the overload capacity of generators, they are able to supply power in
excess of their nominal rated capacity for brief periods of time. Therefore,
large loads which persist only for short durations (such as those due to the
example, may occur due to the failure of other generators within the EPGDS
194
3. Modern aircraft feature APUs with in-flight restart capability, often extending
across the entire flight envelope. Factoring in the APU generator capacity would
doing so would require that the APU be operable in order to dispatch the
aircraft, which is not the case for current commercial aircraft [56]
Based on the above considerations, the required EDG capacity for the whole
In Eq. 145, Pedg,0 is the total generation capacity of all baseline EDGs. In other words,
if the baseline aircraft has two EDGs each rated at 120 kVA, then Pedg,0 = 240 kVA.
∆Pedg,nst is the required increment in total EDG capacity in order to support electrical
loads within a MEA that are not short-term loads (nst: non-short-term). These are
the loads that may persist for extended periods of time that are longer than the
permitted overload duration of the EDGs. ∆Pedg,st is the required increment in total
the aircraft may be lost. This is accounted for the capacity ratio κ, defined as the
Pedg,av
κ= (146)
Pedg
The overload ratio ǫ of a generator (EDG or APUG) is defined as the ratio of the
power that it can deliver during temporary overload conditions to the rated nominal
power output. Let λ = 1 and λ = 0 denote respectively the cases where the APU
generator capacity is and is not factored into the EDG capacity determination.2
2
A default setting of λ = 0 (no credit assumed for APU generator capacity) was used for all
analyses presented in this dissertation, except some presented in Chapter 8 where the effect of λ = 1
(full APU credit) on EPGDS sizing was analyzed.
195
Considering the presence of non-short-term loads Pnst and short-term loads Pst
over the course of the mission, the total electrical load may be expressed as
where the second argument κ indicates that the magnitudes of the loads may depend
on the capacity ratio κ. This allows the modeling of degraded operation modes
in which some non-essential loads may be shed. The required EDG capacity is
1. The available power generation capacity for a given capacity ratio must be
sufficient to meet the maximum non-short-term loads corresponding to that
2. The available power capacity for a given capacity ratio with generators working
(including both non-short-term and short-term loads) for that capacity ratio
subsequent power balance for the second condition yields the increment ∆Pedg,st as
satisfy both short-term and non-short-term loads corresponding to this capacity ratio
196
Repeating the above procedure for multiple failure scenarios with different
capacity ratios κ and corresponding load profiles Pnst (t, κ) and Pst (t, κ) allows the
identification of the most constraining sizing case for the generators. Within the
1. Nominal flight condition: In this case, all EDGs are operational, and therefore
failure scenario is more constraining than the failure of a single generator, since
the latter removes the generating capacity of all generators on the failed engine.
With the required increments in generator capacity known, the total increase in
∆Pedg + ∆Papug
∆Mgen = , (150)
(P/M)gen
where (P/M)gen is the assumed power-to-mass ratio of generators. The nominal value
of (P/M)gen = 2.8 kVA/kg [146, 93] was varied in sensitivity analyses (Chapter 8).
The mass addition due to the incorporation of ATRUs, which are responsible for
conversion of AC power to DC power, is estimated using the peak DC power and the
197
EPGDS template of Fig. 32, these are seen to be electrified actuation functions,
electrified ECS, and ETS (if installed). A fixed value of ηatru = 0.97 [147] is
used for all analyses. The power-to-mass ratio, which is set to a nominal value of
(P/M)atru = 1.54 kW/kg [147], was varied during sensitivity analyses (Chapter 8).
In order to estimate the mass of power distribution elements, the dependency of the
mass on nominal system voltage, voltage drops, power transmitted, and length is first
analyzed. The current I flowing through a conductor is related to the voltage Vnom
L
∆V = IR = Iρe , (153)
Acs
where R is the electrical resistance, ρe the electrical resistivity, and Acs the
cross-sectional area. Combining these two relationships, the cross-sectional area may
be expressed as
Pin ρe
Acs = . (154)
Vnom (∆V /L)
It follows that if ρm is the material density of the conducting cable, then its mass
Mcbl is given by
ρm ρe
Mcbl (Pin , L) = ρm Acs L= Pin L = Kcbl Pin L, (155)
Vnom (∆V /L)
in which (∆V /L) may be thought of as the maximum permissible voltage drop per
unit length and the constant Kcbl can be shown to have units of kg/(VA.m) (kilogram
per volt-ampere per meter). It is clear from Eq. 155 that the mass of a conductor
carrying a given power over a given length can be reduced by (i) using a material of low
198
density and low electrical resistivity, (ii) using a high nominal transmission voltage3 ,
and (iii) allowing a higher permissible voltage drop per unit length. Further, replacing
Considering a conducting cable that runs for a length ℓ0 from a source to a common
distribution point and then splits into two cables that run lengths ℓ1 and ℓ2 to two
junction law yields that the current through the common length is I0 = I1 + I2 .
relationship of Eq. 156, the total mass M of this cable layout may be computed as
′ ′ ′
M = Kcbl I0 ℓ0 + Kcbl I1 ℓ1 + Kcbl I2 ℓ2
′
= Kcbl {(I1 + I2 ) ℓ0 + I1 ℓ1 + I2 ℓ2 }
′
= Kcbl {I1 (ℓ0 + ℓ1 ) + I2 (ℓ0 + ℓ2 )}
′ ′
= Kcbl I1 L1 + Kcbl I2 L2 (157)
The above result (equivalent to that of Eq. 139 for HPGDS) establishes that even
if dedicated cabling does not exist from a source to each consumer, it is sufficient to
consider the direct (non-exclusive) length from a source to a consumer for the purpose
of computing cable mass. The direct lengths of the power distribution elements are
automatically computed from the 3-D geometric model of the aircraft (using Eq. 136)
the necessary mass and electrical properties of conducting materials, nominal system
voltage, and permissible voltage drops. However, attempting to compute the mass
3
Hence the recent trend towards the use of higher voltage networks for aircraft
199
of power distribution elements in this manner would be overly simplistic due to the
following reasons:
1. This estimate would be valid for a single conducting wire, whereas in aircraft
electrical networks bundles of wires are used. For example, a 3-phase AC system
with a neutral would have four wires, a 3-phase AC system with no neutral
2. The estimate would not account for the weight of insulating material. In aircraft
Christou et al. [148] investigated the choice of optimal voltages for MEA wiring
systems, in which the effect of voltage level on the necessary insulation thickness
and the multiplicity of wires within a bundle for different AC and DC systems were
explicitly accounted for. The authors attempted to optimize a wiring system while
considering maximum power transfer and wiring weight with constraints related to
the occurrence of discharges at high voltages. As part of their results, they presented
the power-to-weight ratio (power transfer capability of the wire per unit mass) as
a function of the voltage rating. This has the units of kVA.m/kg (kilo volt-ampere
230VAC
meter per kg), the inverse of the units for Kcbl . Based on their results, 1/Kcbl =
±270VDC
64.6 kVA.m/kg and 1/Kcbl = 86.8 kVA.m/kg were identified for the 230 VAC
and ±270 VDC networks considered. It is important to note that this method of
selecting Kcbl does allow the effect of wire bundles and insulation to be accounted for
The actual wiring length is likely to be somewhat greater than the computed
wiring length due to the need to avoid obstacles whose positions become defined only
later in the design. Also, the mass estimation method does not directly account for
200
the mass of connectors and other fittings required in a wiring system. Therefore, an
installation factor κinst is used to modify the cabling mass estimation:
It is clear from the above relationship that the mass of cabling is subject to uncertainty
as a result of possible uncertainties associated with each of the four terms on the
right-hand side. The impact of this at the architecture level is assessed subsequently
Although the EPGDS does not directly consume power, there is power dissipation as
a result of component inefficiencies and losses in the distribution system. There is,
magnitude of power dissipated will in general increase with the magnitude of power
processed by the EPGDS, and therefore is likely to be more substantial as the degree
of subsystem electrification increases.
The total power dissipation is essentially the difference between the shaft-power
input to the EPGDS generators and the total power delivered to the power consuming
P P
subsystems in either AC or DC form, respectively Pac (t) and Pdc (t). Assuming
an overall efficiency of η (i) for a component, the following relationships exist between
the power input, power output, and power dissipation (loss) of that component:
(i)
(i) Pout (i) (i) (i) (i) 1
η = (i) , Ploss = Pin − Pout = Pout −1 (159)
Pin η (i)
As per the assumed template for the EPGDS (Fig. 32), shaft-power entering the
EPGDS control volume encounters the following inefficiencies en route to the AC and
DC power consumers:
1. Between generator input and AC power consumers: The generators with overall
efficiency ηgen , the power feeders with overall efficiency ηf dr , and the AC
201
distribution system (cabling/wiring) with assumed overall efficiency ηac
2. Between generator input and DC power consumers: The generators and feeders
listed above, the ATRU with overall efficiency ηatru , and the DC distribution
components starting from the power consumers and working upstream to the
generator input, it is easy to see that the total dissipation of the EPGDS is given by
X X
1 1
PEP GDS (t) = −1 Pac (t) + −1 Pdc (t)
ηgen ηf dr ηac ηgen ηf dr ηatru ηdc
(160)
The component efficiencies in the above relationship are summarized in Table 23 and
were identified from [147] and the permissible voltage drops listed in [149]. For a
distribution system with nominal source voltage Vns and a permissible voltage drop
With reference to the PPGDS architecture shown in Fig. 33 and Fig. 34, the
analysis control volume of the PPGDS is considered to include the Precooler Heat
Exchangers (PHXs) and all pneumatic ducting originating at either the PHX exit or
202
the engine bleed ports and terminating at the locations of the pneumatic consumers.
In case a subsystem architecture contains no pneumatic users, the PPGDS is deleted
entirely. The mass estimation for the PHXs and pneumatic ducting is addressed in
There may be significant variation in the temperature of the engine bleed air
depending on the engine throttle setting, the stages from which the bleed air is
extracted, and the port selection logic employed. The function of the PHX is to
down-regulate the bleed air to a pre-defined target temperature. The PHX is typically
an air-to-air cross-flow heat exchanger that is integrated into the engine nacelle and
uses air extracted from the fan stage as the cooling stream. The mass flow rate of
the fan air is modulated by the FACV (Fig. 33) in response to feedback from a
sensor that measures the temperature of the down-regulated air exiting the PHX.
The exit temperature is typically limited to 200 ◦ C, with high temperature and low
The sizing condition for the PHX is the case where the maximum amount of
thermal energy has to be removed from the hot (bleed) stream. Therefore, it
may be hypothesized that this would occur during a combination of the following:
(i) high throttle settings that lead to high temperatures for bleed air entering the
PHX, (ii) high bleed air demands from consuming subsystems, and (iii) additional
demands imposed by failures of other pneumatic channels (either engine or bleed
performance that were presented in [56], in which the takeoff condition with failure
was identified as the constraining sizing case. Based on the above, a simple
relationship for estimating PHX mass is used for twin-engines and four-engined
203
aircraft, subject to the assumption that the following remain invariant with aircraft
size: (i) the heat exchanger technology level, (ii) the inlet temperatures of the hot
and cold streams in the sizing case, and (iii) the target exit temperature of bleed air
in which the summation is applied to all consumers of precooled bleed air, and the
denominator neng /2 ensures that the consumers’ bleed requirements can be satisfied
with only half of the total pneumatic generation capacity available subsequent to
failures. The constant κphx was identified by solving Eq. 161 for κphx for the case of
a known precooler weight [152]. This yielded a value of κphx = 17.33 kg/(kg/s).
Due to the nature of the relationship shown in Eq. 161, the estimated PHX mass
air reduces. This may occur if a consumer’s bleed air requirement reduces (e.g., if
In addition to the fuel consumption penalty imposed by the extraction of bleed air
from the compressor stages, an additional penalty is imposed due to the extraction
of fan air for the purpose of down-regulating the temperature of this bleed air.
Both these effects may be taken into consideration while computing the overall fuel
The mass of pneumatic ducting is proportional to the duct length, but is also
influenced by the mass flow rate of air through the duct, which influences the duct
internal diameter. In addition to this, the fact that the air flowing through the duct
these dependencies but also allow an estimate of the duct mass with knowledge of
204
only a limited number of parameters, the approach described subsequently is taken.
From simple geometry considerations, the mass per unit length M/L of a
thin-walled pneumatic duct of inner diameter Di and wall thickness tw which is made
M π π
=ρ (Di + tw )2 − Di2 = ρ tw (tw + 2Di ). (162)
L 4 4
Using the continuity equation and the equation of state, the duct internal diameter
Di can be related to the mass flow rate of air ṁ of air flowing through the duct as
pnom π 2
ṁ = ρa Acs vmax = D vmax ,
Ra Tnom 4 i
s
4 Ra Tnom
=⇒ Di = ṁ (163)
π pnom vmax
where Ra is the gas constant for air, Tnom is the nominal temperature of the air flowing
through the duct, pnom the nominal internal pressure, and vmax is the maximum
The wall thickness tw may be determined based on the permissible tensile stress
pmax Di pmax Di
tw = = , (164)
2 {σt − pmax (1 − Y )} 2σef f
where a value of Y = 0.4 is used for non-ferrous metals [153]. The wall thickness
relationship above originally uses the outer diameter Do , but here it has been modified
Substituting Eq. 163 and Eq. 164 into Eq. 162 yields the mass of a duct of length
where κinst is an installation factor that accounts for (i) the weight of fittings and
attachments that were not explicitly accounted for and (ii) for the fact that the
205
Table 24: Summary of constant parameters for duct mass estimation
actual ducting length is likely to be longer than the computed one due to installation
parameters occurring in Eq. 165 is shown in Table 24. While the actual pressure
and temperature of bleed air flowing through the ducts varies with time and location,
κduct is determined using nominal values [151]. The pressure pmax used for determining
the wall thickness tw is set in accordance with FAR §25.1438 which requires that the
integrity of ductwork be verified at burst pressures equal to three times the normal
working pressure. Both Grade 2 and Grade 3 titanium are found in aircraft pneumatic
systems [154], and σt is set to the mean value of their permissible tensile stresses at
The fuel consumption penalty as a result of bleed air extraction depends on the mass
flow rate of bleed air extracted and its thermodynamic state at the point of extraction.
Since it is assumed that there is no leakage within the PPGDS, there is no direct bleed
air consumption attributable to the PPGDS. The mass flow rate of bleed air entering
the PPGDS analysis control volume also exits that control volume to supply the
pneumatic power consumers. Thus, the bleed air fuel penalties are accounted for
under the pneumatic power consuming subsystems and not the PPGDS.
206
5.5 Mechanical Power Generation and Distribution
System (MPGDS)
The selected control volume for the MPGDS only contains the accessory gearbox.
The modeling of this subsystem is limited to (i) the estimation of mass changes of
the gearbox with shaft-power demand and (ii) accounting for the mechanical losses
from power transmission through the gearbox. For the same technological SOTA, the
NASA using actual gearbox weight data [102], where the gearbox mass is a function
The horsepower (hp) in Eq. 166 is determined based on the peak shaft-power
requirement accounting for both electrical power and hydraulic power (if any).
The input and output RPMs are set based on data provided in engine type
lower rotational speeds, which results in larger torques applied to the gearbox.
The power dissipation of the MPGDS due to mechanical losses within the gearbox is
estimated as
X
1 X
PM P GDS (t) = −1 Pelec (t) + Phyd (t) (167)
ηgb
207
in which the gearbox efficiency is assumed to be ηgb = 0.97 for a modern
high-performance gearbox [147]. Since the summation of shaft-power requirements
Pspx (t) is performed over all subsystems consuming or dissipating hydraulic or electric
power, the dissipation as computed above is for the entire MPGDS which includes
multiple gearboxes.
The modeling approaches for the Hydraulic Power Generation and Distribution
For each, the sizes and masses of the major components are determined based on
Chapter 4) and relevant sizing scenarios. For the HPGDS, EPGDS, and MPGDS, the
losses is also accounted for. These quantities are fed back into the aircraft sizing and
mission performance analysis to evaluate the impact of the subsystem architecture
208
CHAPTER VI
This chapter addresses Research Question 1 and tests Hypothesis 1, both of which
vehicle and mission level using computationally inexpensive sizing and analysis
methods that require only limited information regarding the aircraft design?
(a) the initial sizes and characteristics of major components in individual power
(b) the initial sizes and characteristics of additional components in power generation
209
(c) the mission performance analysis reflects the direct and indirect effects of the
subsystem architectures on the performance of the propulsion system
Two scenarios are studied in this context. The first (§6.1) analyzes the
subsystem (chosen to be the ECS). The second (§6.2) analyzes the aircraft-level and
For the purpose of conducting these two analyses, the impact of shaft-power and
bleed air extraction on the engine fuel consumption (fuel flow rate) are modeled using
simple relationships that provide first approximations of the magnitude of the off-take
penalties. The incremental fuel flow ∆ẇf,spx due to total shaft-power extraction of
of Nop,eng engines which are assumed to contribute equally to the total shaft-power
Pspx . The constant kp∗ was given as kp∗ = 0.0094 N/W as an average of the penalties
computed at flight altitudes of 0 ft, 10,000 ft, 20,000 ft, and 35,000 ft at Mach numbers
The incremental fuel flow penalty ∆ẇf,bx due to bleed air extraction ẇbld = ṁbld g
per engine is computed following the method of SAE AIR 1168/8 [136] as
Ttet [◦ R] ẇbld
∆ẇf,bx = 0.0335 , (per engine) (169)
2000 Nop,eng
in which Ttet is the turbine entry temperature, for which this dissertation uses a
210
Off-take penalty relationships such as Eq. 168 and Eq. 169 are useful since they
require little additional information other than the time variation of the shaft-power
and bleed air off-takes. On the other hand, the limitation of these relationships is that
these are general ones that do not apply to any one engine in particular and do not
explicitly account for the variation of the magnitude of off-take penalties with flight
condition (altitude, Mach number, and engine power setting). Therefore, they are
suitable for providing approximate indications of the fuel penalties due to off-takes,
but are sufficient for pursuing the intent of this chapter.
To verify whether the developed integrated sizing and analysis environment can
changes to the Pneumatic Power Generation and Distribution System (PPGDS) and
changes to the subsystems architecture and also to the nature of secondary power
off-takes, which are summarized below:
of Cabin Air Compressors (CACs), motors, and power electronics into the ECS.
Further, EPGDS mass increases due to higher rating electrical generators and
2. Reduction of mass: The mass of the PPGDS reduces due to the elimination or
down-sizing of pneumatic ducting.
211
3. Reduction in bleed air off-take: The bleed air off-takes from the engine are
significantly reduced. Due to this, the engine TSFC reduces
shaft-power extraction from the engine, since the electric ECS must be provided
5. Increase in ram drag: The air supplied to the cabin must be admitted through
ram air inlets instead of being bled from the engine, which results in an increase
in ram drag
The overall consequence of electrification of the ECS depends upon the relative
using (i) subsystem-level metrics such as the masses of the ECS, EPGDS, and
PPGDS, (ii) aircraft-level metrics such as the Operating Empty Weight (OEW), and
(iii) mission-level metrics such as the block fuel or the ramp weight.
Using the Total Fuel Impact (TFI), the net impact on mission fuel consumption
may be further decomposed and attributed to the individual causal factors (system
mass, secondary power requirements, and drag increments) of the subsystems affected
(in this case ECS, EPGDS, and PPGDS). In the subsequent sections, these analyses
The ECS, EPGDS, and PPGDS architectures are established within the Subsystem
For SSA-00000, the conventional baseline with pneumatic ECS, the PPGDS
architecture is shown in Fig. 36. Bleed air is extracted from both main engines and
ducted through wing ducts into the pneumatic manifold in the fuselage area. The
pneumatic system is divided into left and right systems, which are interconnected
212
Figure 36: Conventional (pneumatic) ECS architecture (SSA-00000)
213
Figure 38: EPGDS architecture for electric ECS (SSA-00010)
through a cross-flow valve so that one system can supply users on the other side in
case of a failure in the other system. Under nominal operating conditions, the left
system supplies the left wing IPS and the left ECS pack. In addition, since the APU
must be able to support ECS operation on the ground, an APU supply duct runs
from the aft fuselage to the central pneumatic manifold1 .
For the electric ECS design, SSA-00010, the heuristic architecting rules initialize
two electrical generators per engine due to the significantly higher electric load. For
subsequently transformed by ATRUs into DC power to supply the ECS. The resulting
EPGDS architecture is shown in Fig. 37 and Fig. 38. Since the APU must still be
able to power the ECS on the ground, an APU feeder runs from the aft fuselage to
the main electrical/electronics bay to supply the AC system. Thus, it is clear that
The overall analysis approach has already been described previously, and only
1
Although the engine starting function is not covered within this dissertation, this same APU
supply duct is used in a pneumatic starting arrangement for starting the first main engine.
214
a brief summary of relevant details for this particular case is presented. For both
the architectures, the ECS, being the power consuming subsystem, is evaluated first.
shaft-power), and ram air requirements over the course of the defined mission (§4.7).
The two affected power generation and distribution subsystems, the PPGDS and
the EPGDS, are evaluated subsequently. In both cases, the lengths of the power
distribution elements are established based on the architecting rules and the geometric
model of the aircraft. The PPGDS sizing module uses ducting lengths and the mass
flow rate requirements of the residual pneumatic power users to estimate the total
mass of ducting and precoolers (§5.4). In a similar manner, the EPGDS sizing module
determines the masses of the electrical cabling, generators, and power transformation
devices based on the electrical power requirements of the ECS and consideration of
mass, secondary power requirements, and ram air requirements becoming available.
This information is then used to re-evaluate the vehicle’s mission performance (as
described in §3.2.5). Finally, the computation of the TFI and TWI of individual
Figure 39 summarizes how the electrification of ECS for the SSA affects
subsystem-level metrics such as the masses of the ECS, PPGDS, and EPGDS, an
aircraft-level metric such as the empty weight, and mission-level metrics such as
ramp weight and block fuel. As evident from Fig. 39, as a result of electrification
of the ECS, there is a 70 % increase in the mass of the ECS itself. This is due
to the addition of the CACs, electric motors, and power electronics. The relative
down-sizing of the wing pneumatic ducts, and the down-sizing of the precoolers. The
215
Figure 39: Summary of ECS electrification effects for SSA (SSA-00000 vs.
SSA-00010)
Table 25: Detailed mass-∆ investigation for ECS, EPGDS, and PPGDS
relative contributions of ducting and precooler masses are shown in Table 25. There
is considerable mass addition to the EPGDS, in order to support the large electrical
demand of the ECS. The mass addition is due to higher higher-capacity generators,
AC-to-DC power conversion equipment, and also feeders and cables. The relative
The mass additions to the ECS and EPGDS exceed the mass reduction of the
PPGDS, and thus overall there is a 2.39 % increase in the OEW of the aircraft.
However, despite the overall increase in aircraft OEW, the electrification of the ECS
results in a predicted fuel burn reduction of 2.44 %. To investigate the relative
contribution of the subsystems to the overall fuel burn reduction, the Total Fuel
216
Table 26: Aircraft and mission-level impact of ECS electrification for SSA
Impact (TFI) and Total Weight Impact (TWI) breakdown may be analyzed, as shown
in Table 26. The quantities in this matrix are expressed in terms of percentage deltas
(%-∆) relative to the following base quantities: (i) The change in fuel penalty of the
MEA due to weight, shaft-power, bleed air, and direct drag are all expressed as a
%-∆ relative to the baseline architecture’s TFI. Thus, for each row, the elements of
the first four columns add up to the element in the fifth column. (ii) The change in
each subsystem’s mass is expressed as a %-∆ relative to the subsystem’s mass in the
conventional architecture. The figures in this column are identical to those in Fig. 39.
It is seen from Table 26 that the increase in mass of the ECS leads to a
corresponding increase in the fuel penalty incurred over the course of the mission due
to this extra mass. Since the cabin supply air is admitted through dedicated ram air
inlets in case of electric ECS, this results in additional direct (ram) drag. However,
the fuel burn reduction due to the elimination of bleed air off-take for the ECS is
greater in magnitude than the fuel burn penalty due to the additional shaft-power
in the TFI of the electric ECS relative to the conventional (pneumatic) ECS. The
predicted fuel savings is of sufficient magnitude to offset the increase in ECS mass
and result in an overall favorable TWI. For the PPGDS and EPGDS, the TFI and
TWI are driven directly by their respective mass deletion and addition.
217
% ∆ SFC and thrust vs. time % ∆ fuel flow vs. time
5 0
% ∆ SFC
% ∆ SFC, % ∆ Thrust
% ∆ Thrust
% ∆ Fuel Flow
−1
0
−2
−5 −3
0 0.5 1 0 0.5 1
0.8 1
0.6
0.9
0.4
Pneumatic ECS 0.8
0.2
Electric ECS
0 0.7
0 0.5 1 0 0.5 1
Normalized Mission Time Normalized Mission Time
Figure 40: Comparison of thrust, TSFC, fuel flow, fuel burn, and vehicle weight over
the course of the mission (SSA-00010 relative to SSA-00000)
SSA-00000 over the entire mission. To obtain more insight into the relative
performance of one architecture against the other, the time evolution of these
parameters over the mission must be examined. This is shown in Fig. 40, from which
it can be seen that the replacement of ECS bleed air requirements with shaft-power
be seen from the normalized vehicle weight comparison that SSA-00010 is heavier
than SSA-00000 throughout the mission. As a result of this extra mass (the result
of increased equipment weight) and additional ram drag for the ECS, the thrust
requirement of SSA-00010 is higher than that of the baseline. The magnitudes of the
increased thrust and the decreased fuel flow are such that overall there is a reduction
in the required fuel flow rate of SSA-00010 which ultimately results in a mission
218
1
0.8
0.6
0.4
0.2
Normalized ECS Flow Requirement
Normalized electric ECS Power Requirement
0
0 0.2 0.4 0.6 0.8 1
Normalized Time
Figure 41: ECS mass flow and power requirements over mission
for the ECS may be made from Fig. 41, from which it is seen that the mass flow
rate requirement (for both pneumatic ECS and electric ECS) reduces with increasing
altitude (due to the pressurization schedule). The mass flow rate is directly related
to the bleed air penalty for the pneumatic ECS. Despite the reduction in the required
mass flow rate, there is an increase in the electrical power consumed by the CACs
with altitude, due to an increase in the required CAC pressure ratio to bring ambient
The results for ECS electrification for the LTA and the VLA are presented in Fig. 42
and Fig. 43 respectively. In these cases, there is an even more significant addition
of mass to the ECS. The total compression power requirement increases with the
minimum required mass flow rate, which in turn is directly proportional to the
number of cabin occupants. The higher compression power requirement results in
correspondingly higher masses for CACs, drive motors, and power electronics.
219
Figure 42: Summary of ECS electrification effects for LTA (LTA-00000 vs.
LTA-00010)
Figure 43: Summary of ECS electrification effects for VLA (VLA-00000 vs.
VLA-00010)
A significantly larger percentage of the baseline PPGDS mass is eliminated for the
LTA and the VLA compared to the SSA. It was established in Chapter 5, Sec. 5.4
that the mass of pneumatic ducting was approximately proportional to the product
of mass flow rate and ducting length (ṁ L). In case of the VLA and LTA, the
P
contribution of the ECS to the summation (ṁ L) is greater, due to the increasing
220
6.2 Comparison of Two Competing Subsystem
Architectures - All Electric vs. Conventional
In §6.1, the case of electrification of one power consuming subsystem (ECS) was
considered, along with the necessary changes to two power generation and distribution
the aircraft subsystem architecture is considered, in which all the subsystems are
1. The wing planform area should be re-sized to maintain the same wing loading
2. The same thrust-to-weight ratio TSL /WT O as the baseline must be maintained
3. The horizontal and vertical stabilizers must be re-sized to maintain the same
For each of the three aircraft, a transition from a fully conventional subsystems
in Fig. 44, in which all %-∆ quantities are expressed relative to the corresponding
quantities for the baseline architecture. It is seen that for the majority of the actuation
functions there is an increase in systems mass. This is due to the assumption that
with the current technological state-of-the-art, electric actuators are heavier than the
hydraulic actuators they replace. The significant increase in mass of the ECS was
already observed in §6.1. For the WIPS and CIPS, electrification involves the removal
of the piccolo tubes of the conventional architecture and the incorporation of heating
elements and associated power electronics for the electrothermal IPS architecture.
221
Figure 44: Summary of effects of conventional to All Electric subsystem architecture
transition for SSA (SSA-73310 relative to SSA-00000)
Table 27: Subsystem-, aircraft-, and mission-level impact of All Electric subsystem
architecture for SSA
222
the EPGDS and a significant increase in its mass (in excess of what was observed
in §6.1). At the same time, the PPGDS and HPGDS are eliminated completely
from the aircraft, resulting in the elimination of significant mass. The net result is
a 2.13 % increase in the empty weight of the aircraft. A similar increase in OEW of
the aircraft was predicted in the MOET final report [49] for an electrified short-range
reference aircraft similar to the SSA considered here. The TFI and TWI analyses
for the subsystems architecture transition are shown in Table 27. For the actuation
subsystems that show a net increase in systems mass, a corresponding increase in the
fuel penalty due to systems mass is observed. For the case of the ECS, WIPS, and
CIPS, a reduction in the total fuel penalty is seen due to the elimination of bleed air
requirements, which results in a favorable TFI and TWI for these subsystems. At
The increase in empty weight on the one hand and the reduction in mission fuel
requirement on the other yield an increase in the ramp weight of the aircraft of
0.27 %. Due to the re-sizing rules described above, this implies that there is an
increase in the wing area Sw and the thrust rating TSL of the same magnitude.
Similar trends are observed for the electrification of the LTA and VLA subsystem
architectures, which are presented in Appendix A (LTA: Fig. 63 and Table 36,
VLA: Fig. 64 and Table 37). For these cases as well, there is an increase in the OEW
of the aircraft due to the architecture transition and a reduction in fuel consumption
driven primarily by the elimination of the bleed air requirement. In case of the LTA
and the VLA, there is a net reduction in the ramp weight of the aircraft, by 0.65 %
and 0.61 % respectively, which means that wing planform area Sw and sea-level rated
thrust TSL were reduced by the same percentage during the re-sizing.
223
6.3 Chapter Summary
The two architecture comparisons described in this chapter demonstrated that the
electrified ECS) and also competing subsystem architectures (e.g., conventional versus
All Electric) on aircraft-level metrics such as empty weight and mission-level metrics
such as fuel burn and ramp weight. This was verified for the three baseline aircraft
considered (SSA, LTA, and VLA). In both cases, following the sub-statements of
requirements flow-down and using information that either exists during early
(b) determining the impact of the subsystem solutions on the affected power
(c) propagating the effect of the masses, secondary power requirements, and drag
obtain updated estimates of mission-level metrics such as fuel burn and vehicle
weight, not just at the end-points of the mission but as a function of time over
Further, for the comparison between two competing subsystem architectures, each
vehicle was re-sized in accordance with a set of re-sizing rules (statement (d) of
Hypothesis 1). In addition to predicting gross fuel burn reductions, the TFI analysis
for each subsystem allowed the relative contributions stemming from changes in
224
subsystem mass, secondary power extraction, and drag increments (associated with
the subsystem or architecture transitions) to be identified. The results address
aircraft, and mission levels but without requiring detailed knowledge that would not
225
CHAPTER VII
The purpose of this chapter is to describe Experiments 2.1, 2.2, and 2.3, which
are aimed at verifying certain capabilities of the integrated sizing and analysis
environment. The benchmarking of the automatic architecture definition algorithm
of the net architecture impact into contributions from masses, secondary power
In §5.1, heuristics for the association of power sources, power systems, and power
existing aircraft. In addition, templates for the EPGDS and PPGDS were also
elements based on (i) the number of engines, (ii) the number of power consumers,
i.e., the number of ailerons, spoilers, etc., (iii) the nature of the secondary power
power sources, power systems, and power consumers determined by the algorithm
226
are passed to successive sizing and analysis modules. The algorithm also generates
a diagrammatic representation of the architecture connectivity in the form of the
Considering the snapshot of Fig. 45 as an example, the top-left corner shows the
nature of off-takes from the prime movers (engines E1, E2, etc. and the APU). If
bleed air is extracted from the engines and APU, that bleed is marked B-E1, B-E2,...,
and B-APU. Similarly, if EDPs and EDGs are run off the engine, these are marked
EDP-E1, G1-E1, etc. To the right of this, the pneumatic system definition is shown.
The left and right halves of the pneumatic system are marked P1 and P2, and are
connected by the Cross-Bleed Valve (CBV). The bleed extracted from the engines and
the APU supplies either system P1 or P2. Below the pneumatic system definition,
the hydraulic system definition is shown. For each hydraulic system, the pumps that
are involved in pressurizing it are shown above the system names H1, H2, and H3.
Additionally, the RAT is shown marked above the system which it supplies in case
the electrical system. As per the electrical system template described in §5.1.2, there
is a separate AC bus for each main engine generator. In case of AC-to-DC power
conversion, the connectivity of the AC systems to the DC systems is also shown (e.g.,
in Fig. 48). Thus, the top part of each snapshot shows the connectivity of prime
movers to power sources, and power sources to power systems. The central part of
each snapshot shows the association of power systems with the power consumers.
and the validity of the generated subsystem architectures (Experiment 2.1), the known
subsystem architectures of some existing architectures are used as references. For each
such case, the number of engines, number of power consumers (ailerons, leading-edge
227
and trailing-edge devices, spoilers, etc.), and the type of secondary power required by
the power consumers is provided as an input. The subsystem architectures generated
by the algorithm (Figs. 45-48) are then compared with the actual architectures. For
cases where the association created by the algorithm differs from the actual association
between the same two elements in the real architecture, the latter is indicated by red
text in parentheses.
The first two benchmarking cases are the four-engined Airbus A340 and the
twin-engined Boeing 777, for both of which the subsystems architecture is largely
conventional. It is seen from Fig. 45 and Fig. 46 that in several cases there
the algorithm and the actual connectivity existing in these two aircraft. However,
this is to be expected since the heuristic for hydraulic system to spoiler association
was a generalized one developed from the inspection of multiple hydraulic system
architectures. Further, in several cases, hydraulic system H1 is predicted in place of
H3 or vice versa. The hydraulic system definition shows that in terms of power source
redundancy, these two systems are essentially equivalent. It is also noteworthy that
the connectivity for the primary flight control surfaces, thrust reversers, and landing
parallel to hydraulic ones for the ailerons, elevators, rudders, and spoilers, and the
electrification of the TRAS. It differs from the actual A380 subsystems architecture
in that the Airbus A380 uses Electrical Backup Hydraulic Actuators (EBHAs) for the
rudder and some of the spoilers, whereas these are not considered within the scope
of this dissertation1 . It is noteworthy that for this case, the algorithm eliminates
1
Within the scope of this dissertation, an actuator is supplied with either hydraulic power or
electric power, but cannot receive both types of power (which is the case for EBHAs)
228
229
Figure 45: Subsystem architecture snapshot generated by heuristic algorithm for Airbus A340 (for cases where the algorithm
predicted a connection different from the actually existing connection, the latter is indicated using red text within parentheses)
230
Figure 46: Subsystem architecture snapshot generated by heuristic algorithm for Boeing 777 (for cases where the algorithm
predicted a connection different from the actually existing connection, the latter is indicated using red text within parentheses)
one of the three hydraulic systems and generates a 2H/2E actuation architecture, as
found on the Airbus A380. Hydraulic supply is removed from the TRAS and replaced
with electric supply. For the ailerons, elevators, and rudders, one hydraulic supply is
removed and replaced with one electric supply. Since this aircraft has a split rudder,
each panel is provided with two actuators, one hydraulic and one electric.
MEA) are provided to the architecture definition algorithm. With regard to the
association of hydraulic systems with spoilers and some of the ailerons, the same
observations that were made for the Airbus A340 and Boeing 777 benchmarking
cases are applicable to this case as well. Due to the greater electrical power demand
caused by the electrification of the ECS, the algorithm initializes two generators per
engine (and also for the APU) instead of the usual one. Due to the electrification
of the ECS and WIPS functions, the pneumatic systems P1 and P2 are deleted.
The conventional hydraulic supplies are removed from the brake and THS actuation
functions and replaced with equivalent electric supplies. Finally, three spoilers per
wing are electrified (in reality, the Boeing 787 has two electrified spoilers per wing).
It is clear from the benchmarking cases that while the algorithm does not precisely
match each and every system-to-load connection within the subsystem architectures,
what is seen in reality for these two aircraft. The conclusion from these observations
is that this algorithm can be used in lieu of a formal FTA/FMEA in order to rapidly
determine (i) the required number of power sources and systems and (ii) feasible
connectivity among subsystem architecture elements based on the subsystem solutions
231
232
Figure 47: Subsystem architecture snapshot generated by heuristic algorithm for Airbus A380
233
Figure 48: Subsystem architecture snapshot generated by heuristic algorithm for Boeing 787 (for cases where the algorithm
predicted a connection different from the actually existing connection, the latter is indicated using red text within parentheses)
7.2 Capturing Effect of Subsystem Architecture on
Propulsion System Sizing and Performance
First approximations for the impact of shaft-power and bleed air extraction on
the engine fuel consumption (fuel flow rate) were presented in Chapter 6. It
was pointed out that relationships of this nature are convenient as they require
little additional information other than the instantaneous shaft-power and bleed
air off-takes. However, these relationships do not account for the fact that the
engine’s sensitivity to shaft-power and bleed air extraction depends on the engine
cycle parameters, which vary from engine to engine even within the same thrust
tool into the integrated sizing and analysis environment may be infeasible from the
truth model is then queried in order to find the incremental fuel flow rate as a result of
shaft-power and bleed air off-takes for a given flight condition. The coefficients thus
obtained are incorporated into the integrated environment and used for architecture
assessments. Within the scope of this dissertation, this method is employed only
for the SSA (Experiments 3.1, 3.2, and Experiment 4.0 with architecture-to-engine
where SSA results for these experiments are presented. For the LTA and VLA, the
the propulsion system truth model and the determination of the fuel flow penalty
234
7.2.1 Propulsion System Truth Model
The propulsion system truth model was created using the Environmental Design
Space (EDS) tool. The EDS was developed under the sponsorship of NASA and the
US Federal Aviation Administration’s Office of Environment and Energy (FAA/AEE)
design modules include CMPGEN for generation of compressor maps, the Numerical
Propulsion System Simulation (NPSS) tool for analysis of the thermodynamic cycle,
and WATE for engine flow path analysis and weight estimation [159]. In EDS, a
number of different design points are used as part of the Multi Design Point (MDP)
1. Aerodynamic Design Point (ADP): Mach 0.80, 35,000 ft, ISA - This is
2. Top of climb: Mach 0.85, 35,000 ft, ISA - At this point, thrust requirements
3. Takeoff: Mach 0.25, sea-level, ISA + 27◦ F - The takeoff thrust requirement
4. Sea-level static, installed, ISA + 27◦ F - The sea-level static thrust requirement
For each of these points, the shaft-power extraction and customer bleed air extraction
power requirements into the sizing of the engine, the shaft-power and bleed air
235
extraction corresponding to a One Engine Inoperative (OEI) scenario (assumed to
occur anywhere in the mission profile) for a particular subsystem architecture are
stated above are then used to size three engines corresponding to the secondary power
1. Mixed Off-take (MO) Engine: Sized taking into account the mixed secondary
2. Reduced Bleed (RB) Engine: Sized taking into account the secondary power
3. Zero Bleed (ZB) Engine: Sized taking into account the pure shaft-power
Since the bleed air requirements are lower for running-wet anti-icing compared to
of pneumatic running-wet WIPS and CIPS. Further, since the electrical power
requirements are lower for running-wet anti-icing relative to evaporative anti-icing and
lower still for de-icing, the ZB engine automatically satisfies the power requirements
the bleed air requirements for ice protection reach their maximum at an intermediate
altitude and become zero at cruise. As a result, the RB engine, sized assuming
electric ECS and pneumatic evaporative IPS, has sufficient shaft-power and bleed
capacity to meet the off-take requirement of all subsystem architectures where the
236
Figure 49: Inputs and outputs for truth model function call (executed as a function
call to NPSS with a MATLAB wrapper)
ECS/WIPS/CIPS group has been only partially electrified (i.e., at least one member
to compute the mass flow rate of air that must be extracted from the fan in order
to cool the bleed air to a target PHX exit temperature (set to a constant 200◦ C).
The extraction of this mass flow rate also imposes a fuel consumption penalty on the
engine. Thus, the TSFC degradation due to bleed air extraction computed using the
truth model accounts for not only the direct penalty due to the extraction of the bleed
air itself, but also the additional penalty due to the extraction of fan air in order to
down-regulate the temperature of the bleed air. A simple bleed port selection logic
settings, bleed air extraction from a higher pressure compressor stage is considered.
For each engine, the truth model is executed for combinations of altitude, Mach
number, and power settings but with zero shaft-power and bleed air demand in order
to generate a basic engine performance data-table (engine deck). This is used in order
to evaluate the fuel consumption for the basic mission (§3.2.5) prior to accounting for
the additional fuel consumption due to subsystem secondary power requirements and
237
By wrapping a MATLAB script around the truth model, a truth model function
call for each of the three engines is created with the inputs and outputs shown in
Fig. 49. For the case of the ZB engine, the bleed air demand is always zero and the
PHX model is not enabled (as the PHX would have been physically eliminated from
bleedless architectures). The truth model is queried by means of this function call in
due to shaft-power and bleed air off-takes, as described in the following section.
the shaft-power extraction [63]. However, such a linear relationship for bleed air
extraction is only true for relatively low mass flow rates of bleed air [63]. Off-line
investigation of the impact of shaft-power and bleed air extraction on the predicted
TSFC of the engine truth model confirmed these facts. Further, a sensitivity analysis
performed on the TSFC revealed a slight interaction effect between the shaft-power
and bleed air extraction. Based on these observations, a fuel flow rate penalty function
∆SF C
λsf c = = K1 ṁ2bx + K2 ṁbx + K3 ṁbx Pspx + K4 Pspx (170)
SF C0
For a given flight condition (altitude, Mach number, and thrust setting), the
coefficients K1 , . . . K4 are determined by querying the truth model in order to solve
ṁ2 ṁbx,1 ṁ2bx,1 Pspx,1 Pspx,1 K SF C1 − SF C0
bx,1 1
ṁ2 2
bx,2 ṁbx,2 ṁbx,2 Pspx,2 Pspx,2 K2 1 SF C2 − SF C0
= (171)
2 SF C0
ṁbx,3 ṁbx,3 ṁ2bx,3 Pspx,3
Pspx,3 K3 SF C3 − SF C0
ṁ2bx,4 ṁbx,4 ṁ2bx,4 Pspx,4 Pspx,4 K4 SF C4 − SF C0
238
It is clear that to solve K = [K1 , . . . , K4 ]T for a single flight condition, five
evaluations of the truth model are required - a first evaluation to establish the
basic TSFC (SF C0 ) without any shaft-power or bleed off-takes, and four subsequent
and B matrices of Eq. 171. In literature, the TSFC degradation due to shaft-power
and bleed air extraction is often expressed as percentage increase in TSFC due to
ṁbx,2 = 1 lb/s, (iii) Pspx,3 = 100 hp, ṁbx,3 = 1 lb/s, and (iv) Pspx,4 = 40 hp, ṁbx,4 =
0.6 lb/s. This is done for all engines designed to provide mixed off-takes (i.e., both
shaft-power and bleed). For the case of a bleedless engine, only the coefficient K4 is
Since this evaluation is for a particular flight condition, the coefficients depend
on altitude, Mach number, and thrust setting, i.e., Ki = Ki (h, M, T ), i = 1, . . . , 4. In
1. The mission performance analysis is set up such that the flight profile given by
t ∈ [0, tf ], h(t), M(t) is nearly invariant for all missions evaluated, regardless of
the basic TSFC through the dependency of engine TSFC on thrust setting.2
2
It should be noted however that variations in fuel flow rate due to variations in thrust are
captured directly through the terms of Eq. 20
239
t
time τ = tf
∈ [0, 1], i.e., K(τ ) = [K1 (τ ), . . . , K4 (τ )]T . Thus, for any mission being
analyzed, the coefficient-set may be obtained from a 1-dimensional look-up.
The fuel consumption increment predicted using the penalty coefficients determined
by the procedure outlined above was tested against that predicted by the truth model
for the same flight condition and secondary power off-takes. To do this, the truth
model is first queried with the altitude, Mach number, thrust, shaft-power off-take,
and bleed air off-take corresponding to each instant of time within a selected mission
profile. The fuel flow rates predicted by the truth model are integrated forward in
time to obtain the truth (or reference) fuel consumption for that mission.
time histories of shaft-power and bleed air extraction, the degradation in TSFC (λsf c ,
Eq. 170) is computed for each instant of time. With the degraded TSFC given
by SF C = SF C0 (1 + λsf c ) and the thrust from the mission profile, the fuel flow
predicted by the interpolated penalty coefficients is computed for each time instant.
These too are integrated forward in time to obtain the predicted fuel consumption
for the mission. This is then compared to the truth (reference) fuel consumption
When the comparison is carried out using the same mission profile that had
been used to determine the coefficients, the error is negligibly small. Across the
sized engines, for a constant ±1,000 ft perturbation to the altitude time history
(i.e., a constant addition or subtraction of 1,000 ft to each altitude data point), the
magnitude of the predicted fuel consumption error is limited to 0.57 %. Similarly, for
a constant ±0.01 perturbation in the Mach number time history, the magnitude of
the error is limited to 0.68 %. For a ±5 % perturbation to the thrust time history,
240
the magnitude of the predicted fuel consumption error is limited to 0.39 %. It should
be noted that since the mission performance analysis is set up such that the flight
profile given by t ∈ [0, tf ], h(t), M(t) is nearly invariant for all missions evaluated,
the actual deviations of altitude and Mach number time history during architecture
The conclusion from the above is that it is feasible to develop a truth model for gas
the truth model can be incorporated into the mission performance analysis without
In order to guide the focus of design refinement efforts for subsystems, the design
must instead be derived by (i) decomposing the net impact of the subsystem
architecture into the contributions from different subsystems and (ii) decomposing
the contribution from each subsystem further into contributions arising from mass,
secondary power requirements, and direct drag increments. The approach followed
Experiment 2.3). The breakdown of the net architecture impact into the contributions
from mass, secondary power requirements, and direct drag increments arising from
efforts for ECS and WIPS solutions are presented in §7.3.2 and §7.3.3. Finally, a
241
feasibility assessment for an optional subsystem such as an Electric Taxiing System
(ETS) for the SSA, LTA, and VLA baselines is presented in §7.3.4.
The approach to evaluating the total fuel impact of the overall subsystem architecture
was discussed in §3.2.5, where a system of equations (Eq. 17) was developed to
account for the time-varying effects of secondary power extraction and direct drag
increments. It was mentioned that evaluating this equation system for individual
permitted the evaluation of all the components of a subsystem’s Total Fuel Impact
The relative magnitudes of the four effects (shaft-power, bleed, drag increment, and
weight) may be represented through the following four ratios:
(i) (i) (i) (i)
∆wf,spx (i) ∆wf,bx (i) ∆wf,d ∆wf,w
κ(i)
spx = , κbx = , κd = , κ(i)
w = (173)
T F I (i) T F I (i) T F I (i) T F I (i)
The TFI itself may be expressed as a percentage of the block fuel (BF). The weight
weight (OEW). Finally, the TWI of each subsystem may expressed as a percentage of
the aircraft’s Maximum Ramp Weight (MRW). This information may be represented
in a tabular form as shown in Table 28, where the breakdown shown corresponds to
is re-computed once per iteration and stored in matrix form for each subsystem
242
1. Level 1 - decomposition of net impact into contributions from each subsystem:
This information may be obtained by inspecting the normalized TFI column
(T F I (i) /BF ) of Table 28. This shows which subsystems within the architecture
have the most dominant effect on fuel consumption. For this architecture
mass, secondary power requirements, and drag increments: For any subsystem,
(i)
this information may be obtained by inspecting the relative magnitudes of κw ,
(i) (i) (i)
κspx , κbx , and κd . The relative magnitudes may be used to guide design
refinement efforts for each subsystem. For instance, it is clear that the impact of
Table 28: Decomposition of net subsystem architecture fuel impact into contributions
from mass, secondary power requirements, and drag increments of individual
subsystems (Breakdown shown for SSA-73310, All Electric subsystems architecture,
PPGDS and HPGDS completely eliminated, ETS not installed)
243
For other subsystems such as the ECS, there are significant contributions from
weight, shaft-power requirement, and drag increment. In such cases, the design
of the subsystem should not focus solely on the minimization of mass. This is
shaft-power or bleed air), and direct drag increments on the Total Fuel Impact
(TFI) of conventional (pneumatic) and electric ECS are shown in Fig. 50 for the
SSA, LTA, and VLA baselines. It is seen that for all three baselines, the necessary
pneumatic off-take (bleed air) accounts for a very significant percentage of the TFI
of the pneumatic ECS. A much smaller percentage of the TFI is accounted for by the
system mass, whose effect is seen to reduce with increase in the aircraft size (while
the contribution of bleed air off-take increases). The remaining small percentage of
Figure 50: Decomposition of total fuel impact for pneumatic and electric ECS into
contributions from mass, secondary power requirement, and direct drag penalty
244
the TFI is accounted for by ram drag caused by the cooling stream required for the
ECS heat exchangers.
proportionately high weighting factor would have to be associated with the reduction
of bleed air penalty, with smaller weighting factors for the minimization of weight
and ram air requirement. The magnitude of the penalty is proportional to the work
done to compress the bleed air prior to extraction, which may be expressed as
where ṁbleed is the mass flow rate of bleed air and hbleed is the specific enthalpy of the
bleed at its extraction point. Thus, the penalty may be reduced by
• Reducing the mass flow rate requirement: This is already seen in current
pneumatic ECS designs, where recirculation is used to reduce the net mass
flow rate of air that has to be extracted from the engines [13]. In fact, in
• Reducing the specific enthalpy: The bleed port selection logic employed in
all modern commercial aircraft essentially attempts to achieve this through
localized energy management. Bleed air is extracted from a lower pressure port
whenever possible unless the temperature and pressure available at that port
is insufficient (due to low power settings). However, even the air bled from the
lower pressure port often has excessive temperature and pressure (as pointed
Thus, while the bleed air penalty constitutes a very large percentage of the TFI
of a pneumatic ECS design, it would appear that sufficient design freedom does not
exist to substantially reduce this penalty. ECS design efforts could focus on reduction
245
of the impact of mass (mass optimization) or improvement of the overall effectiveness
of heat exchange (pack performance optimization, which would reduce the ram air
requirement). However, as evident from Fig. 50, these collectively account for a much
smaller fraction of the TFI of the pneumatic ECS. Therefore, the gains arising out of
such design refinement efforts would likely not be of very large magnitudes.
Figure 50 also shows that the scenario is quite different for the electric ECS
solution, where mass, secondary power requirement, and drag increments have a
more balanced contribution. The following observations may be made regarding the
• Mass: This has a far more substantial contribution to the TFI. As mentioned
due to the incorporation of CACs, electric motors, and power electronics. Due
CACs, which increases with increasing mass flow rate and higher discharge
pressure requirements. Similar to the case of pneumatic ECS, the mass flow
required discharge pressure may be reduced by using ECS packs that are
• Ram drag: This has two sources, the first being the cabin supply air which
is admitted through dedicated ram air inlets, and the second being the cooling
stream for the pack heat exchangers. Since minimum cabin airflow requirements
have to be met, the first source of drag can only be reduced up to a limit. The
246
The conclusion from these observations is that greater design freedom exists for
refining the design of an electric ECS solution, either through design improvements
SOTA of components such as motors and power electronics. On the other hand,
the design freedom for a conventional (pneumatic) ECS is more limited. The
permits the identification of the relative contributions of ECS mass, secondary power
requirements, and direct drag penalties to the mission-level fuel consumption. These
for ECS design optimization. The focus of the optimization is seen to differ based on
The contributions from mass, secondary power requirement, and direct drag penalty
for the WIPS solutions that were considered in this dissertation are shown in Fig. 51
for the SSA baseline. The trends for the LTA and VLA baseline are similar, but in
those cases the contribution of the WIPS mass to its TFI is higher since those aircraft
It is seen from Fig. 51 that the penalty due to bleed air off-take is the most
significant for the pneumatic evaporative anti-icing system. In this case, higher
bleed air mass flow rate is required in order to maintain the surface at the higher
lower, since in this case, the surface only has to be maintained at a temperature
sufficient to prevent impinging water from freezing. This reduces the bleed air
requirement, and in this case, bleed air and system mass have roughly comparable
247
Figure 51: Decomposition of total fuel impact for different WIPS solutions into
contributions from mass, secondary power requirement, and direct drag penalty
The contribution of system mass towards the WIPS TFI is much higher for the
case of the electrothermal WIPS solutions. Based on the modeling approach followed
in this dissertation, the masses of the heating elements themselves are identical for
the mass of the power control electronics, which are assumed to be proportional to
the peak power requirement of the WIPS, are therefore higher for the case of the
anti-icing system. For the same reason, the contribution from shaft-power off-take
requirements is also higher for the electrothermal running-wet anti-icing system. For
the de-icing system, there is also a direct drag contribution due to the buildup of ice
during the heat-off period of the system’s cycle. The drag due to ice buildup is a
function of the heater-off time, which coincides with the ice buildup time. Thus, for
a de-icing system, there is a tradeoff between the penalty from the electrical power
requirement and the penalty from the drag increment due to ice buildup.
248
7.3.4 Feasibility Analysis for Electric Taxiing System (ETS)
always requires addition of hardware to the aircraft and thus causes an increase
in the OEW. The direct mass additions are due to the electric motors, gearboxes,
and power electronics for the ETS, and due to higher-capacity APU generators and
additional cabling within the EPGDS. Further, with the modeling approach followed
in this dissertation, the LGAS actuators are re-sized if the incorporation of an ETS
is detected, since the ETS changes the main landing gear actuating moment due to
the addition of considerable mass at the lowest portion of the landing gear leg. As
seen from Fig. 24 and Eq. 80, the mass addition and power requirement increase
significantly with aircraft mass, leading to the general belief that ETS is suitable only
for smaller aircraft flying short-haul routes. This belief was tested by analyzing the
feasibility of an ETS for all three baselines - SSA, LTA, and VLA.
The intended purpose of the ETS is to save fuel, and the fuel savings (if any)
depend on the relative magnitudes of fuel saved on the ground while taxiing and
extra fuel consumed in-flight due to the added mass of the ETS. It was shown in
§4.6.3 that the fuel savings directly depend on the time that the ETS is operational
on the ground (which is the total taxiing time less the time for main engine warm-up
and cool-down), and the flight distance (or duration). Since these vary significantly,
the feasibility assessment of the ETS is not a single-point analysis. Therefore, for
each baseline, the assessment is made for a range of trip distances and taxiing times.
The results for the SSA, LTA, and VLA are summarized in Fig. 52 and Fig. 53. For
each trip distance, the percentage variation in fuel consumption is computed with
respect to the fuel consumption of a conventional baseline with no ETS flying the
same distance.
For the SSA, it is seen from Fig. 52 that for total taxiing times in excess of 15
minutes, the ETS permits a net reduction in fuel consumption over a substantial
249
E−Taxi Feasibility Analysis −SSA
2
−1
% ∆ Fuel
−2
−3
portion of the range capability of the aircraft. Further, for total taxiing times in
excess of 20 minutes, a fuel burn advantage is retained even at 100 % range. These
predictions are in reasonable agreement with estimates provided by Airbus for the
case of electric taxiing applied to the Airbus A320 aircraft [32]. The feasibility
analysis suggests that ETS may be viable for smaller commercial aircraft that spend
a significant amount of time taxiing, and which fly numerous short-haul flights at
On the other hand, the feasibility assessment for the LTA and VLA (Fig. 53)
supports the claim that ETS is not suitable for aircraft of these sizes. In this case,
the in-flight fuel penalty due to mass addition nullifies the fuel saved even for the
longest taxiing times considered at only 45-60 % of the aircrafts’ range capabilities
(for shorter taxiing times, the cross-over point is even earlier). These aircraft are
typically operated on long-haul trans-oceanic routes, where they are regularly flown
250
E−Taxi Feasibility Analysis −LTA
2
−1
% ∆ Fuel
−2
−3
−1
% ∆ Fuel
−2
−3
Figure 53: Feasibility assessment of electric taxiing for LTA and VLA
251
out to much larger percentages of their maximum range capability than the SSA
typically is. The overall fuel penalty incurred for such operations would therefore
This assessment assumes that the incorporation of ETS is the only modification
to the aircraft, i.e., all other subsystems remain conventional. If other subsystems
are also electrified and have a larger favorable impact, then it is possible that
the resulting architecture may in certain cases show some savings relative to the
conventional baseline even in the operating areas of Fig. 53 where the sole inclusion
of ETS results in increased fuel consumption. This in itself does not mean that
ETS should necessarily be considered practical for larger aircraft (LTA and VLA)
provided other subsystems are being electrified, since it could be counter-argued that
incorporating ETS simply serves to negate some of the fuel saving potential of the
other electrified subsystems. Nevertheless, since the stated goals of this dissertation
include analyzing various combinatorial possibilities for subsystems and the effect of
is considered for all three baselines and not just for the SSA baseline alone.
This chapter presented experiments that verified certain capabilities of the integrated
sizing and analysis environment such as (i) the automatic definition of subsystem
power requirements, and drag increments in order to determine the focus of subsystem
design efforts. The results presented and the observations made from them support
252
CHAPTER VIII
It has been mentioned previously that to be suitable for a conceptual design stage
analysis, the methods used to size and analyze the subsystems must allow for rapid
evaluation and require only limited information regarding the design. As a result, a
other cases, models for novel subsystem architecture components were created and
used even though there was not a large volume of reference data available with which
to validate or calibrate these models. This uncertainty regarding the accuracy of the
SOTA corresponding to a future time. However, other than basing such projections on
253
curve for these components is much steeper than those for components within
a conventional architecture, which have already reached the point of technology
sensitivity analyses are performed, and these are discussed in §8.2. For these analyses,
the secondary power extraction penalties are computed using the higher-fidelity
may be associated with model parameters that are known to be the sources of the
chosen to represent major sources of uncertainty in the model. The overall sensitivities
of top-level metrics such as block fuel (BF), maximum ramp weight (MRW), and
Considerable uncertainty exists regarding the magnitude of the actuation loads for the
actuation functions (FCAS, LGAS, NWSS, WBS, and TRAS). For the primary flight
controls, the actuation loads are derived from an estimate of the hinge moments.
254
These depend on the definition of the control surface relative to the main lifting
surface, however the control surface dimensions are finalized later in the design
process. Similarly, the identification of the flight conditions yielding the maximum
hinge moment for these control surfaces requires estimation of key speeds such as
maneuver speed, design cruise speed, and design dive speed, and certain stability
and control coefficients. These too are not known with complete certainty during
early design. The magnitude of the actuating loads for the high-lift devices varies
considerably depending on the flap mechanism kinematics [66], which are not finalized
Similar uncertainty exists for the actuation loads for subsystems associated with
the landing gear group. Landing gear actuation loads are affected by both the estimate
of the landing gear leg mass (which is also affected by the estimates of the masses of
the associated subsystems such as the brakes and the steering mechanism) and details
regarding the mechanism kinematics. These affect the sizing and thus the mass of
the LGAS actuators. The NWSS actuation load depends on several key geometric
characteristics such as the rake angle and the tire trail, for which only estimates from
preliminary design guidelines are available. A similar argument also applies for the
case of brake actuation loads, which affect the flow requirements of hydraulic brakes
and the mass and power requirements of electric brakes. TRAS actuation power
requirements are inferred from a very limited set of data, and are also uncertain.
actuation loads described above are under-predicted, while others are over-predicted.
K-factor is associated with each and every actuation load, the overall sensitivity of
the architecture’s performance to actuation load predictions may also be assessed by
the use of a single K-factor for all actuation loads. This K-factor Kact−loads , which is
255
of a multiplicative nature, is assumed to modify either actuation loads (when these
are computed directly) or actuation power requirements.
While the activities for the LGAS, NWSS, WBS, TRAS, and the high-lift devices
of the FCAS are reasonably well defined in terms of when in the flight they occur
and for what duration, the same is not the case for the activities of the primary flight
control surfaces and spoilers of the FCAS. The amplitudes of their excursions are
such activity (as a function of total mission time) is also dependent upon atmospheric
conditions. For example, while cruising under calm atmospheric conditions, there may
be very little control surface movement, while considerable control surface movements
may be required to stabilize the aircraft’s trajectory and attitude under gusty or
turbulent conditions. To account for these uncertainties, two additional K-factors are
defined for the FCAS. The first factor Kf cas−amp multiplies with the peak amplitude
of assumed control surface excursions δpeak (Eq. 44 of §4.1.9), while the second factor
For the ECS analysis, the heat transfer rate through the fuselage walls is computed
based on an estimate of the total heat transfer area and the effective thermal resistance
across the fuselage skin. The trend seen in modern aircraft is towards the increased
use of composite materials for the aircraft structure. The thermal conductivity of
composites is typically less than that of metallic skin, which has the effect of reducing
the rate of heat transfer across the fuselage wall for the same temperature gradient.
At the same time, the contributors to the internal heat load, such as the galley and
IFE loads, show an increasing trend in modern aircraft. Further, the effect of direct
solar radiation incidence on the fuselage and solar radiation through transparencies
are not accounted for in the current analysis. The net effect of each of the factors
mentioned above is to increase the net heat load that the ECS has to remove from
256
the cabin in order to maintain it at the desired temperature, which is modeled using
a single multiplicative K-factor Kecs−load to multiply the Q̇int term of Eq. 94 (§4.7).
For the ice protection systems (WIPS and CIPS), there is uncertainty regarding
(i) the estimate of the surface area for which ice protection is required, (ii) the estimate
of the heat flux required to provide ice protection to a protected surface for a given
set of flight and atmospheric conditions, and (iii) the assumed overall efficiency of
the IPS layout for both pneumatic and electrothermal IPS setups. The ultimate
consequence of (i)-(iii) is uncertainty regarding the net heat rate of the WIPS and
CIPS, which affects the required bleed air mass flow rate for pneumatic IPS and the
required electric power demand for electrothermal IPS. As with the actuation loads,
it is not possible to say a priori whether the analysis approach employed will result
possible to under-predict the WIPS loads while over-predicting the CIPS loads, or
vice versa. In order to analyze the variation of responses subjected to this uncertainty,
a single multiplicative K-factor Kips−load is linked to the computed IPS bleed air and
form Mcbl (Pin , L) = κinst Kcbl Pin L. Variations in the computed mass may occur due
to (i) variations in mass properties and electrical properties of the conducting material
(affecting Kcbl ), (ii) variations in cabling length L due to the need to avoid obstacles,
(iii) variations in installation factor κinst caused by attachments and fittings. These
are captured through a multiplicative K-factor Kcbl−wt which multiplies the computed
is used to multiply the computed mass of each duct within the PPGDS to account
for mass variations due to the same causes.
The K-factors discussed above and their ranges are summarized in Table 29. The
257
Table 29: Summary of model sensitivity parameters
K-factor Subsystem(s) affected Range
Kact−loads FCAS, LGAS, NWSS, WBS, TRAS [0.75, 1.25]
Kf cas−amp FCAS [0.50, 2.00]
Kf cas−acti FCAS [0.50, 2.00]
Kecs−load ECS [1.00, 2.00]
Kips−load WIPS, CIPS [0.50, 2.00]
Kcbl−wt EPGDS [0.75, 1.25]
Kduct−wt PPGDS [0.75, 1.25]
ranges for each factor were determined through engineering discretion and based
on the perceived accuracy of the modeling approaches that are utilized. Instead of
performing the sensitivity analysis for all 512 subsystem architectures of Table 4, the
following architectures are down-selected for the SSA, LTA, and VLA (Note: the
For each of the subsystem architectures chosen for analysis, a full-factorial set
of K-factor combinations is considered, with each K-factor permitted two levels
corresponding to the upper and lower bounds shown in Table 29. This leads to
27 = 128 case evaluations for each of the four subsystem architectures for each of the
258
8.1.2 Assessment of Architecture Sensitivities
Regardless of the aircraft size (SSA, LTA, or VLA), many similarities were noted in
text, while those for the LTA and VLA are presented in Appendix B.1.
epistemic uncertainty parameters for the case of the SSA are shown in Fig. 54.
The percentage changes in block fuel (BF), maximum ramp weight (MRW), and
operating empty weight (OEW) are all computed relative to the conventional baseline
with all K-factors set to unity. The dots represent the four architectures with all
K-factors set to unity. For K-factor values other than unity, the position of the
architecture on the plots shifts. The individual shifted positions corresponding to all
the K-factor settings are not shown. Instead, the dotted lines represent the convex
hull of the region occupied by the shifted architectures. It is clear that the most
desirable location for any architecture on the BF-MRW and BF-OEW plots is in the
lower-left quadrant (indicating lower BF, RW, and OEW relative to the baseline).
possible to have architectures in which the RW and OEW are higher than those for
the baseline, but the BF is less than that of the baseline. This situation corresponds
to the upper-left quadrant of one or both plots. For architectures falling in the
remaining two quadrants (lower-right and upper-right), the BF is higher than that of
Since the baseline itself is affected by the K-factors listed in Table 29, the
Fig. 54. The original position of the all-electric actuation architecture, SSA-70000,
259
∆ Block Fuel vs. ∆ Ramp Weight ∆ Block Fuel vs. ∆ Op. Empty Weight
1.5
3 SSA−03310
0.5 1
0 0
SSA−73310
−0.5 SSA−73310 −1
−1 −2
−3 SSA−70000
−1.5 SSA−70000 SSA−00000 SSA−00000
−2 −1 0 1 2 −2 −1 0 1 2
% ∆ Block Fuel % ∆ Block Fuel
of the overlap, the only conclusion that can be drawn from these plots is that the
of SSA-00000 for its worst case scenario. Also, the improvement for SSA-70000 for
the best-case scenario is more than that of SSA-00000 for its best-case scenario.
Inspection of the convex hulls for architectures SSA-03310 and SSA-73310 reveals
that there are no quadrant-crossings for these two architectures due to the variation
of the epistemic uncertainty K-factors. Thus, a fuel burn advantage is predicted for
these two architectures (relative to the baseline SSA-00000 with all K-factors set to
unity) over the entire range of variation of the K-factors. Between SSA-03310 and
Table 29 on the BF, OEW, and MRW from inspection of Fig. 54 alone. For this. it
is necessary to inspect Fig. 55, which shows the sensitivities of BF, OEW, and MRW
260
Figure 55: Impact of individual epistemic uncertainty K-factors on performance of
SSA-00000, SSA-70000, SSA-03310, and SSA-73310
261
of the SSA architectures to the individual K-factors (corresponding sensitivities for
the LTA and VLA are presented in Appendix B.1). In this plot, in which the plotting
ranges for BF, OEW, and MRW are identical for all four architectures, the sensitivity
slope of the corresponding line. The following observations may be made regarding
the effects that the K-factors have on each of the architectures analyzed:
1. Kact−loads : The sensitivity of BF, RW, and OEW to this K-factor depends on
whether the actuation functions are all-hydraulic or all-electric. It is seen that
fact that electric actuator power-to-mass and force-to-mass ratios are assumed
SOTA. Therefore, the same variation in the magnitude of the actuation load
control surfaces (thus larger number of actuators in total) and higher actuation
2. Kf cas−amp and Kf cas−acti : The sensitivity of BF, RW, and OEW to these two
K-factors is seen to be marginal for all three aircraft. The sensitivity of BF is
only marginally higher for the case of hydraulic actuation. This is due to the
fact that hydraulic actuation power requirements are proportional to the rates of
the control surfaces (Eq. 48), whose mission-averaged value increases with both
these K-factors. The electric actuation power requirement, as seen from Eq. 47,
is a more complex function depending not only on excursion amplitude but also
on the rate at which hinge moments change with unit control surface deflection.
The conclusion that may be drawn is that while control surface duty cycles
affect actuator design significantly (more so for electric actuators [161, 162]),
the impact on top-level parameters (BF, OEW, and MRW) is not as significant
262
3. Kecs−load : This K-factor has no effect on OEW for any of the SSA and LTA
architectures. This indicates that for these two aircraft, even though the net
an increase in the mass flow rate (in other words, the lower bound for cabin
inlet temperature is not reached). Such an increase in mass flow rate would
CACs, motors, and power electronics being required. As the mass flow rate
requirement does not change, neither do the secondary power requirements of
both pneumatic ECS and electric ECS. Therefore, the change in BF is caused by
an increase in ram drag due to increased ram air flow requirement across the ECS
pack heat exchangers. The effect is seen to be slightly more pronounced for the
case of electric ECS, indicating that the increase in cooling airflow requirement
is higher in this case. However, since ECS pack optimization was not attempted
for either the conventional or electric ECS, no further conclusion can be drawn
from this observation. For the VLA, the increased ECS load does result in the
lower bound for cabin inlet temperature being reached. As a result, the mass
flow rate of supply air has to be increased, resulting in increased penalty from
bleed air and shaft-power extraction for the VLA architectures with pneumatic
the masses of the heater mats of the electrothermal IPS are computed on a
mass-per-unit-area basis and those of the piccolo tubes of the pneumatic IPS on
a mass-per-unit-length basis, the masses of the IPS do not change in either case
due to the variation in Kips−load . Variations in IPS heat load clearly affect the
bleed air requirement of the pneumatic IPS and the electric power requirement
of the electrothermal IPS. However, since the IPS operates for a reasonably small
263
fraction of the flight, the observed effect on block fuel is only partially due to the
increased off-takes. The majority of the effect is due to the variations in OEW
of the aircraft. For pneumatic IPS, this is due to the variations of PPGDS duct
and precooler sizes based on the IPS bleed air requirements. For electrothermal
IPS, this is due to variations in EPGDS cable and generator masses caused by
two architectures. Since the pneumatic system is removed for SSA-03310 and
SSA-73310, the effect of this K-factor is naturally nil. The effect of Kcbl−wt
with vehicle size, presumably due to the fact that the net ducting and cable
with which to assess the effect of assumed or projected technological SOTA on the
For the actuation functions (FCAS, LGAS, NWSS, WBS, TRAS), electric actuators
are currently assessed as being heavier than their conventional hydraulic counterparts.
With the progression of technology, the power-to-mass or force-to-mass ratios of the
264
a technology K-factor directly to such figures of merit. This is due to the fact that not
all sub-components of an electric actuator are likely to see improvements of the same
order of magnitude. For example, the main components of an EHA are the electric
motor and power electronics, a hydraulic pump (and perhaps an accumulator), and a
are traditional hydraulic components whose design has already reached a high degree
components of an EMA are the electric motor and power electronics, a gearbox, and
a ballscrew. The gearbox and the ballscrew are traditional mechanical components
which have been similarly refined over a long period. Thus, mass reductions for these
actuators (and other electric actuators as well) would likely stem primarily from
motors and power electronics with higher power-to-mass ratios. Therefore, it is more
logical to associate the K-factors with the motor and power electronics power-to-mass
ratios than with the overall figure of merit of the actuators themselves. It was this
thought process, in fact, which motivated the generic electric actuator mass estimation
relationship that was developed and discussed previously (Chapter 4, Eq. 31).
Motors are not merely limited to electrified actuation subsystems, but also to
major subsystems like electric ECS, where the CACs are driven by electric motors.
Many considerations go into the selection of the type of motor for a particular
application (even different types of actuators may employ different types of motors),
necessarily be of the same magnitude for each motor design. However, in the early
design phases, there may be insufficient information available regarding the type of
motor most suitable for a given application. The interest lies in determining or
bounding the extent to which improvements in electric motor technology may affect
265
a single K-factor Kem−sota is used “across-the-board” to represent the technology
improvement for all electric motors for all subsystems.
individual loads will require power regulation prior to use [142]. Similar to electric
motors, significant differences in design exist for power conversion equipment, and
it is also unlikely that the improvements in power-to-mass ratio will be uniform for
all such designs. However, for the same reason stated for electric motors, a single
(thus significant mass), part of which comes from the mass of the CACs. Therefore,
a K-factor Kcac−sota is applied to assess the effect of improvements in technological
SOTA for this component. It is not clear whether improvements in the power-to-mass
ratio of CACs can be of the same magnitude as improvements for motors and power
electronics. However, similar ranges were set for Kcac−sota as for Kem−sota and Kpe−sota .
electrification of the major aircraft subsystems, the mass of the generators (EDGs and
APUGs) also increases significantly. Therefore, improvements in the technological
SOTA (power-to-mass ratio) of generators would clearly have an effect on the net
addition of mass to the EPGDS. This effect is modeled using the K-factor Kgen−sota .
A final K-factor Kapu−start is used to model the effect of factoring in the power
generation capacity of the APU generators into the sizing of the EDGs. Therefore,
this is not a technology K-factor in the strict sense of the word, since no improvements
in APU technology are being modeled using it. Instead, a setting of Kapu−start = 0
simply indicates that APU generator capacity cannot be factored into the sizing of the
266
Table 30: Summary of technological SOTA K-factors
K-factor Subsystem(s) affected Settings
Kem−sota actuation, ECS {1,2,3}
Kpe−sota actuation, ECS, EPGDS {1,2,3}
Kcac−sota ECS {1,2,3}
Kgen−sota EPGDS {1,2,3}
Kapu−start EPGDS {0,1}
EPGDS, while Kapu−start = 1 assumes that the EDG ratings can be down-sized based
Table 30 summarizes the technology K-factors discussed above along with the
discrete settings of each for which the architectures were assessed. In all cases, the
setting of K(.) = 1 for the technology K-factors indicates current technological SOTA,
while settings K(.) = 2 and K(.) = 3 indicate respectively a two-fold and three-fold
improvement in the SOTA. The impacts of technology SOTA on the performance of
the following SSA, LTA, and VLA subsystem architectures are assessed:
technology K-factors is shown in Fig. 56. Similar figures for the LTA and VLA are
267
∆ Block Fuel vs. ∆ Ramp Weight ∆ Block Fuel vs. ∆ Op. Empty Weight
SSA−73311
SSA−73311 3
−2
−1
−3
−2 0 2 −2 0 2
% ∆ Block Fuel % ∆ Block Fuel
presented in Appendix B.2. For each architecture, the marker represents the location
of the design with all technology K-factors set to unity, while the line emanating
from each marker shows the direction of movement of the design when the technology
Since the variations in the technology K-factors lead to the reduction of masses
of motors, power electronics, CACs and generators, the net impact is a reduction in
vehicle mass. Thus, the lines seen in the plots are the sensitivities of the MRW and
the BF to changes in the OEW. That the relationship is approximately linear can
be established starting with a manipulation of the Breguet range equation (for cruise
flight only) to yield the fuel burn penalty due to an increment in OEW:
ct R/(V (L/D))
∆BF = ∆OEW e − 1 = E ∆OEW (175)
From the above, it is easy to show that the three %-∆ quantities being plotted are
268
Thus, it is established that %-∆ BF and %-∆ MRW are linearly related to %-∆ OEW.
The magnitude of %-∆ OEW is greater for the more heavily electrified architectures
(in which motors, power electronics, CACs, and generators account for a higher
percentage of the OEW), and therefore, it follows that %-∆ BF and %-∆ MRW
magnitudes are greater for those architectures as well. As a result, for each aircraft
size (SSA, LTA, and VLA), the net displacement of the architecture on the plot
and SSA-73311, between which the only difference is the inclusion of the Electric
Taxiing System (ETS). With all technology K-factors set to unity, SSA-73311 has a
slight fuel burn penalty relative to SSA-73310. This is due to the fact that the extra
fuel consumed transporting the mass of the e-Taxi system over the mission exceeds the
mass of fuel saved while taxiing on the ground with the ETS for the assumed taxiing
duration. However, with Kem−sota = Kpe−sota = 3, there is a substantial reduction
in the mass penalty associated with the ETS motors and power electronics. In that
case, SSA-73311 has a slight fuel burn advantage relative to SSA-73310. However,
this is not the case for the LTA and VLA (Appendix B.2), for which the installation
The individual impacts of the technology K-factors on the BF, OEW, and MRW
may be analyzed by inspection of Fig. 57, which is for the SSA. Similar figures for
the LTA and VLA are presented in Appendix B.2, since the significant trends are
similar for all three aircraft sizes. In these plots, the plotting ranges for BF, OEW,
and MRW are identical for all four architectures. Therefore, the sensitivity of a
269
Figure 57: Impact of individual technological SOTA K-factors on performance of
SSA-70000, SSA-03310, SSA-73310, and SSA-73311
270
1. Kem−sota : The impact of motor power-to-mass ratio (SOTA) increases as the
total contribution of motors to OEW increases. Therefore, it is quite prominent
for SSA-73310, which has electric motors for all actuation functions as well as
to drive the ECS CACs. However, the effect on SSA-73311 is greater still, due
architecture. Within the analysis volume for SSA-70000, these include the power
converters for the actuators and also the ATRUs which convert AC power to DC
power. However, the effect is greater for SSA-03310 due to the high rating of
the CAC power electronics and the increased rating of the ATRUs. It is greater
still for SSA-73310 which combines electric actuation with electric ECS. Finally,
the impact on SSA-73311 is greatest due to the additional power electronics for
the ETS. Similar trends are also seen for the LTA and VLA (Appendix B.2)
compressor is feasible
the required total capacity of the EPGDS increases, and is thus highest for the
due to the additional APU generator capacity required for the ETS and ground
operation of the ECS, and additional EDG capacity to support all electrified
subsystems in-flight
5. Kapu−start : Taking the power generation capability of the APU generators into
account allows the main engine generators to be sized to a lower power rating,
271
thus saving weight. This has a direct effect on the OEW, which subsequently
affects the BF and MRW. The magnitude of the weight savings is directly related
to the power rating of the APU generators, and is greatest for architectures in
which the ECS is electric and/or an ETS is installed. In these cases, the APU
However, it should be noted that there are other practical considerations that
may also determine whether it is feasible to consider the APU generator capacity
while sizing the main engine generators. First, this would require a guarantee
that the APU can be re-started at any point in the flight envelope. Second, it
may not be possible to dispatch the airplane with an inoperative APU generator
Question 3. This was performed through the judicious association of a limited number
magnitude of the advantage but not its existence. This was not the case for
architectures that had only a marginal advantage relative to the baseline, in which
case the continued existence of such an advantage could not be guaranteed. When
observed for subsystem architectures where these components were present in higher
capacities. The fact that trends and sensitivities of this nature can be detected and
272
CHAPTER IX
of all architectures within the architectural design space defined in Table 4 (Chapter 3)
are analyzed for the SSA, LTA, and VLA (§9.1). Additionally, the effect of factoring in
the subsystem architecture secondary power requirements into the sizing and selection
of the engine is shown as a special case for the SSA (§9.2). Finally, a case study is
presented in which it is assumed that electric and hydraulic actuators sized for the
same actuation requirements have identical mass (§9.3).
All 512 subsystem architecture combinations present within the Matrix of Alternatives
presented in Table 4 (Chapter 3) are evaluated for the SSA, LTA, and VLA. For
these evaluations, the impacts of secondary power off-takes are represented using the
simplified method with constant penalty coefficients (as introduced in Chapter 6).
In each case, the vehicles are re-sized to the same wing loading, thrust-to-weight
ratio, and tail volume coefficients as the baseline (regardless of whether re-sizing
where the Block Fuel (BF), Operating Empty Weight (OEW), and Maximum Ramp
Weight (MRW) of the architectures are shown as %-∆ quantities relative to the
corresponding quantities for the baseline with conventional subsystem architecture.
273
(a) Relative performance of 512 subsystem architectures for SSA
Figure 58: Comparison of 512 subsystem architectures for SSA, LTA, and VLA
274
As a result, these baselines (SSA-00000, LTA-00000, and VLA-00000) occupy the
origins of the respective plots. Inspection of Fig. 58 shows that for all three vehicle
sizes, the architecture combinations fall within one of four clusters. Inspection of the
subsystem architectures of points within each cluster reveals that cluster affiliation
is determined by the subsystem solution for the ECS (pneumatic or electric) and the
1. Cluster 1 (Pneumatic ECS, no ETS installed): Designs within this cluster are
in the vicinity of the origin. Some of the designs show improvements in fuel
consumption relative to the baseline, while others show a degradation. For the
SSA, the majority of the designs within this cluster show improvement in fuel
2. Cluster 2 (Electric ECS, no ETS installed): Designs within this cluster show
for the VLA. For the LTA and the VLA, designs within this cluster also show
reductions in MRW. For all three vehicles, architectures within this cluster have
3. Cluster 3 (Electric ECS, ETS installed): Designs within this cluster show an
increase in both OEW and MRW relative to the baseline. For the SSA, this
cluster is offset almost directly upwards relative to Cluster 2. For the LTA
and VLA however, the offset is upward and to the right. Therefore, for each
point within Cluster 3 for the LTA and VLA, the fuel burn performance is
worse compared to the corresponding point in Cluster 2 which lacks the ETS.
Essentially, this indicates that regardless of the solutions employed for other
subsystems, the inclusion of ETS for the LTA and VLA is impractical
4. Cluster 4 (Pneumatic ECS, ETS installed): All points within this cluster are
heavier in terms of OEW and MRW. For the LTA and VLA, designs within this
275
cluster consume more fuel than the baseline. This is simply further proof that
the incorporation of ETS is infeasible for vehicles of this size (this was also seen
It is clear therefore that the nature of the ECS (pneumatic or electric) and the
architecture lies in. Within a particular cluster, the variations in BF, OEW, and
MRW occur due to variations in the designs of the actuation and ice protection
subsystems. This is analyzed further for Clusters 1 and 2 (since Clusters 3 and
4 are infeasible/impractical for the LTA and VLA). The relative impacts of the
actuation and IPS architectures within Clusters 1 and 2 for the SSA, LTA, and
VLA are summarized in Table 31, 32, and 33 respectively. For each aircraft, a
reference architecture is chosen for each cluster. For Cluster 1, the fully conventional
For Cluster 2, an architecture where electrification was limited to the ECS is chosen
as the reference (SSA-00010, LTA-00010, and VLA-00010).
electric actuators relative to hydraulic ones. For the SSA, the net result of these
trend is reversed for the LTA and the VLA, in which case there is an increase in
the OEW. This is due to a combination of higher actuation loads leading to a
greater surface controls mass penalty for electric actuators, longer and heavier
electrical cables and feeders, and a larger increase in required generator capacity.
276
of the counter-acting effects, the influence of actuation architecture on the BF
performance of the LTA is negligible. For the VLA, the OEW increase results
in a degradation in BF performance
IPS architecture, for most of the combinations of WIPS and CIPS solutions
Table 31: Architecture impacts within Clusters 1 and 2 for SSA (Note: All %-∆
quantities are with respect to the conventional architecture SSA-00000. All ∆(%-∆)
quantities are with respect to the reference architecture within each Cluster)
Cluster 1 Cluster 2
Reference architecture SSA-00000 SSA-00010
(pneumatic ECS) (electric ECS)
%-∆ BF 0.00 - 2.47
%-∆ MRW 0.00 +0.62
%-∆ OEW 0.00 +2.38
Cluster characterization
%-∆ BF 0.00+0.06
−0.79 - 2.47+0.01
−0.84
277
WIPS, CIPS, PPGDS, and EPGDS). Slight degradations in BF performance
occur for the electrothermal evaporative CIPS design. In this case, the high
control equipment, and higher gage electrical cabling. These mass additions
from the CIPS and PPGDS. The magnitude of BF improvements reduces with
aircraft size, being highest for the SSA and lowest for the VLA
The conclusion from this analysis is that the ECS design (pneumatic or electric)
Table 32: Architecture impacts within Clusters 1 and 2 for LTA (Note: All %-∆
quantities are with respect to the conventional architecture LTA-00000. All ∆(%-∆)
quantities are with respect to the reference architecture within each Cluster )
Cluster 1 Cluster 2
Reference architecture LTA-00000 LTA-00010
(pneumatic ECS) (electric ECS)
%-∆ BF 0.00 - 2.73
%-∆ MRW 0.00 - 0.50
%-∆ OEW 0.00 +1.67
Cluster characterization
%-∆ BF 0.00+0.15
−0.36 - 2.73+0.06
−0.40
278
solutions employed for the remaining subsystems have an impact that is of a smaller
magnitude. This is evident from the clustering of subsystem architectures that is
seen in Fig. 58. The impact of ECS electrification on the subsystem architecture
the effect that the exchange of bleed air requirement for shaft-power requirement has
The ten best-performing subsystem architectures for the SSA, LTA, and VLA are
listed in Table 34. For convenience, the architecture identification scheme is presented
again in Fig. 59. For a given aircraft size, the ten best-performing architectures fall
Table 33: Architecture impacts within Clusters 1 and 2 for VLA (Note: All %-∆
quantities are with respect to the conventional architecture VLA-00000. All ∆(%-∆)
quantities are with respect to the reference architecture within each Cluster )
Cluster 1 Cluster 2
Reference architecture VLA-00000 VLA-00010
(pneumatic ECS) (electric ECS)
%-∆ BF 0.00 - 4.20
%-∆ MRW 0.00 - 1.18
%-∆ OEW 0.00 +1.44
Cluster characterization
%-∆ BF 0.00+0.57
−0.22 - 4.20+0.36
−0.25
279
Table 34: Best-performing SSA, LTA, and VLA subsystem architectures with respect
to fuel consumption improvement relative to conventional subsystem architecture
within a very narrow range of fuel consumption improvement. Thus, even though
relative to the baseline conventional architecture, the intent of this tabulation is not
280
1. Electric Taxiing System (ETS) - For reasons already stated previously, ETS
does not feature in any of the top architectures for the LTA and the VLA. For
the SSA however, architectures containing ETS do make appearances. For all
such architectures, a counterpart not containing ETS is also present in the list
for the SSA, LTA, and VLA contain electric ECS. Referring to Fig. 58, all
these architectures lie within Cluster 2 for the LTA and VLA, and within either
Cluster 2 or Cluster 3 for the SSA. The dominant effect of ECS electrification
avoid the use of pneumatic evaporative ice protection due to the associated
bleed air penalty (Fig. 51). There is only one exception: VLA-03010, which
pneumatic running-wet anti-icing systems. For the SSA and LTA, in which
4. Actuation functions: For the SSA and LTA, the presence of actuation Packages
and thrust reversers). Due to the high actuation loads for the VLA and
the assumed technological SOTA (whereby electric actuators are heavier than
281
due to electrification is more significant for the VLA than for the LTA and SSA.
At the same time, higher actuation loads result in higher power requirements,
which increases the mass added to the electrical system (EPGDS) through
the VLA, these mass additions outweigh the progressive mass deletion due to
commonalities exist: the top architectures for all three aircraft sizes are either
(through electrification of ECS, electrification of either the WIPS or the CIPS, and
the use of a running-wet design for the remaining pneumatic ice protection function).
In this case the 512 architectures for the SSA are evaluated using the scheduled
coefficients (§7.2) to represent the penalties due to shaft-power and bleed air
extraction. Such coefficients were computed off-line for three engines sized for three
different shaft-power and bleed air extraction limits, as shown in Table 35. Also
shown in Table 35 are the criteria for associating the three engines with the 512
subsystem architectures. As per this association logic, the Mixed Off-take (MO)
engine is used for all architectures where the ECS, WIPS, and CIPS are pneumatic
(conventional). The Zero Bleed (ZB) engine is used for all bleedless architectures,
in which the ECS, WIPS, and CIPS are electrified. As a result, the shaft-power
capability of the ZB engine is significantly higher than that of the MO engine, while
the customer bleed availability is eliminated. For all architectures where at least one
or more from among the ECS, WIPS, and CIPS remain pneumatic, the Reduced Bleed
282
Table 35: Secondary power extraction limits for Mixed Off-take (MO), Reduced
Bleed (RB), and Zero Bleed (ZB) engines for SSA thrust class and association rules
between engines and ECS, WIPS, and CIPS architectures
(RB) engine is used. To be able to meet the shaft-power and bleed air requirements
Fig. 60. Once more, the architectures are seen to appear in clusters. However, the
cluster characteristics are somewhat different from and more complex than those seen
previously (in Fig. 58). These new SSA clusters are described below in brief:
Mixed Off-take (MO) engine. The performance variations within this cluster
are caused by the changes in actuation architecture and the effect of reduced
As such the fuel consumption benefits of designs within this cluster are limited
designs use the Reduced Bleed (RB) engine. The fuel burn reduction of
designs within this cluster is driven by the electrification of the ECS. As noted
283
Figure 60: Performance comparison of SSA subsystem architectures using scheduled
penalty coefficients for secondary power off-takes and the architecture-to-engine
association logic of Table 35. Absence or presence of ETS from designs within clusters
is indicated using solid and dotted bounding lines respectively
previously, ECS electrification adds significant mass to both the ECS and the
These designs also use the Reduced Bleed (RB) engine. The fuel burn
reduction for designs within this cluster are driven by the partial reduction
or complete elimination of the IPS bleed air requirement and also by reduced
OEW on account on mass deletions from the PPGDS due to WIPS and/or
CIPS electrification. This cluster shows the greatest fuel burn reduction among
284
4. Cluster 4 (Bleedless architecture, ZB Engine): These are bleedless architectures
where the ECS, WIPS, and CIPS are all electrified. Designs within this cluster
use the Zero Bleed (ZB) engine. These designs show a reasonable fuel burn
reduction relative to the baseline, but not as large as those shown by designs
Given the prior observations from Fig. 58 (for which the analysis did not consider
off-takes were factored into the sizing of the RB and ZB engines, and therefore
basic TSFC (accounting for only propulsive power requirements and not
latter.
using ZB engine, Cluster 4), as shown in Fig. 61(a). Despite the basic TSFC
observation made above, it is seen that the overall TSFC of SSA-03300 is higher
for the majority of the flight (due to the pneumatic ECS bleed requirement).
However, SSA-03310 is heavier than SSA-03300 owing to the mass additions
285
SSA−03300 (pneu. ECS, elec. IPS, RB−Eng) relative to SSA−03310 (bleedless arch., ZB−Eng)
2 1
%−∆ SFC %−∆ Vehicle Weight
1.5 %−∆ Thrust %−∆ Fuel Burn
0.5
%−∆ Fuel Flow
1
0
0.5
0 −0.5
−0.5
−1
−1
−1.5
−1.5
−2 −2
0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1
Normalized Time Normalized Time
SSA−03300 (pneu. ECS, elec. IPS, RB−Eng) relative to SSA−00010 (elec ECS, RB−Eng)
2 1
%−∆ SFC %−∆ Vehicle Weight
1.5 %−∆ Thrust %−∆ Fuel Burn
0.5
%−∆ Fuel Flow
1
0
0.5
0 −0.5
−0.5
−1
−1
−1.5
−1.5
−2 −2
0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 1
Normalized Time Normalized Time
to the ECS and EPGDS due to ECS electrification and has additional ECS
ram drag. Thus, SSA-03300 has a lower thrust requirement, and the two
counter-acting effects of thrust and TSFC result in a lower fuel flow rate for
though the former retain pneumatic ECS: This is analyzed by comparing the
mission performance of SSA-03300 (electrified WIPS and CIPS, pneumatic
286
architecture also using RB engine, Cluster 2), as shown in Fig. 61(b). The
observations and reasoning are identical to those above. SSA-03300 has a higher
TSFC, but a lower thrust requirement that ultimately leads to lower fuel burn
This analysis demonstrates that logic for associating the subsystem architecture
with one of several engines sized for different secondary power off-take limits can
In the prior analysis (summarized in Fig. 58), the electrification of the ECS had a
dominant effect on the performance of the subsystem architecture, with the design of
penalty coefficients for bleed extraction identified for the MO and RB engines using
the NPSS engine truth model are of a smaller magnitude than the constant coefficient
obtained from the SAE AIR 1168/8 method [163] (Chapter 6). At the same time,
the predicted penalty for shaft-power extraction is higher than that predicted by the
kp∗ method of Scholz (Chapter 6). The net result is a reduction in the off-take fuel
not bleedless. On the other hand, an engine optimized for reduced (but non-zero)
bleed extraction and higher shaft-power extraction may offer acceptable performance
for a number of MEA subsystem architectures. Since cost and time constraints
would make it infeasible to design a bespoke engine for each More Electric subsystem
287
9.3 Case Study: Equal Electric and Hydraulic Actuator
Masses
is generally assumed that electric actuators will weigh more than hydraulic actuators
designed for the same actuation requirements [73]. This reduces the competitiveness
actuation loads, as evident from the absence of such architectures for the VLA from
Table 34. In Chapter 8, the impact of advancement in electric motor and power
that the electric actuators (or actuation system) for the FCAS, LGAS, NWSS, and
TRAS weigh exactly the same as the hydraulic actuators (or actuation systems) that
they replace. This assumption however is not applied to the WBS, since electric
brake actuation will require the incorporation of EMAs which will always add some
mass to the WBS. With this assumption in place, the progressive electrification of
1. Mass deletions from the 5,000 psi HPGDS due to elimination of progressively
SSA, LTA, and VLA are presented in Fig. 62 in the form of %-∆ OEW, %-∆ BF,
and %-∆ MRW relative to the conventional actuation architecture. The following
1. For the SSA, the performance improvement is the most pronounced and is also
288
Figure 62: Effect of electrification of actuation functions for SSA, LTA, and VLA
under the technology assumption that electric actuators for FCAS, LGAS, NWSS,
and TRAS weigh exactly the same as the hydraulic actuators they replace
monotonic. Even though there is a mass addition to the WBS for Package 1
(electrification of WBS and TRAS), the mass deletion from the HPGDS is of
2. For the LTA, Package 1 results in a slight OEW penalty due to the mass addition
caused by electrification of the WBS. However, subsequent actuation packages
289
3. For the VLA, the mass addition due to the electrification of the WBS in
Package 1 is more significant. Although subsequent packages see monotonic
Thus, it is evident that even with this optimistic technology assumption, complete
approximately 0.42 % for the SSA, and approximately 0.32 % for the LTA and VLA.
While these savings are not negligible, it must be noted that (i) they are subject to a
very optimistic technology assumption (that electric actuators suffer no mass penalty
relative to hydraulic ones), and (ii) the reduction in fuel consumption is almost an
subsystem architectures for three aircraft sizes. It was found that architectures fell
into one of four distinct clusters depending on whether or not the ECS was electrified
and whether or not an Electric Taxiing System (ETS) was installed on the aircraft.
The presence of ETS within architectures on the two larger aircraft (LTA and VLA)
was also observed in a previous analysis (Experiment 2.3), and as such it is possible
Weight (OEW) but also a reduction in the fuel consumption. LTA and VLA (but not
SSA) architectures in which the ECS was electrified (and which did not feature an
ETS) showed a reduction in Maximum Ramp Weight (MRW), implying that these
designs benefited from the down-sizing that occurred during the re-sizing iterations.
ECS electrification was found to have a dominant effect on the performance of any
290
subsystem architecture featuring it. The solutions for all other subsystems were
seen to have impacts of a significantly lower magnitude. When the SSA architecture
space was evaluated employing scheduled penalty coefficients derived from querying a
of the architecture clusters was seen to change. While electric ECS architectures still
time-history of mission performance parameters revealed that this was largely due to
the significant mass addition associated with ECS electrification.
The best performing architectures for the SSA and LTA differed from those for the
SSA and LTA architectures featured actuation packages with significant electrification
of the actuation subsystems. This was not the case for the VLA, due to the
significant mass additions to the actuation subsystems themselves and to the EPGDS
due to the higher actuation loads and actuation power requirements. Even when
electric actuators were assumed to have no mass penalty relative to hydraulic ones,
it was found that the potential for performance improvement was still limited to a
5,000 psi state-of-the-art (rather than 3,000 psi) for the conventional hydraulics.
These observations are of course subject to the assumed characteristics of the
the combinatorial sub-space for actuation architectures. The three aircraft sizes
did however show a similarity in that the best-performing architectures were either
bleedless ones or ones that had significantly reduced bleed air requirements. In the
291
In conclusion, the developed integrated sizing and analysis environment
allowed Research Question 2 to be addressed and yielded insight regarding the
with variation in the aircraft size. Since the performance evaluation for all the
supported. The results presented in this chapter and the preceding ones allowed
the major research questions to be addressed and the hypotheses to be tested. The
final chapter of this dissertation summarizes the core contributions of this work, the
major conclusions drawn, and the avenues identified for further extensions of the
methodology.
292
CHAPTER X
This chapter concludes the dissertation with a summary of the key contributions, a
10.1 Contributions
integrated and modular approach for the simultaneous sizing of an aircraft and its
subsystems in the early design phase. In particular, it allows rapid assessments of the
effect of novel subsystem architectures for which little or no historical data exists.
A core aim of the modeling approach was to ensure that the subsystem models
developed for sizing and analysis were suitable for use during conceptual or early
design phases. This aim was decomposed into two main requirements: (i) that the
number of subsystem architectures within a reasonable amount of time, and (ii) that
the models require as input only information that would either be available or easily
estimated during the early design phases. A salient feature of the demonstrated
approach is that the required input from the analyst is limited to the following:
293
3. Settings for K-factors that modify internal model parameters to simulate
epistemic uncertainty or advancement in technological state-of-the-art (only for
Further, unlike some other tools and approaches, no input from the analyst is
required for determining the connectivity existing among elements of the subsystem
evaluated successively without any input from the analyst and (ii) it does not require
that the analyst be a Subject Matter Expert (SME) in order to carry out meaningful
The approach to the sizing and evaluation of the aircraft subsystems revolves
around a generic representation of the subsystem architecture featuring (i) prime
movers, (ii) power sources, (iii) power systems, (iv) power distribution elements, and
Subsystems responsible for the generation, distribution, and consumption of the three
main types of secondary power (pneumatic, hydraulic, and electric) were viewed as
special cases of this generic framework. The significance of this representation is the
fact that it can accommodate all onboard subsystems, including those not directly
metrics of interest, the following standard information from each subsystem was
propagated upward to the aircraft-level and mission-level: (i) subsystem mass,
294
(ii) shaft-power requirement, if any, (iii) bleed air requirement, if any, and (iv) direct
drag increment, if any. The latter three are evaluated by the relevant subsystem
modules and may be arbitrary functions of time. Unlike some other approaches,
assumed constant over a flight phase or over any extended interval of time.
level decomposes the net impact into the contributions from individual subsystems.
The second level decomposes the contribution of each subsystem into component
and drag increments. These constituent impacts may be used as rational weighting
between aircraft sizing and subsystem sizing. Once the impact of the subsystem
of re-sizing rules, the aircraft geometry and thrust are re-sized based on impacts
originating from the subsystem-level, following which the subsystems are once more
sized based on the updated aircraft-level information. Due to the parametric
definitions of control surfaces and ice protection zones relative to a main lifting surface
such as a wing (to give an example), the re-sizing at the aircraft level automatically
three aircraft of varying sizes were considered: a Small Single-aisle Aircraft (SSA),
a Large Twin-aisle Aircraft (LTA), and a Very Large Aircraft (VLA). These were
chosen since they collectively span a wide spectrum of gross weight. For each of
295
these aircraft, a subset of the subsystem architecture space was down-selected based
on certain physical design constraints for subsystems and subject to an assumption
actuation functions. All candidate architectures within the identified subset were
evaluated exhaustively. Such an approach made the analyses tractable but at the
same time permitted the evaluation of a sufficiently large number of More Electric
Aircraft subsystem architectures. Most importantly, it provided insight into the effect
of vehicle size on subsystem architecture performance. A further notable contribution
arising out of the presented approach is the fact that it facilitates the investigation
of subsystem architectures.
created using a higher-fidelity propulsion system analysis tool was used as a truth
model. The truth model was queried to find penalty coefficients representing the
effect of shaft-power and bleed air extraction on the engine fuel consumption. These
coefficients were used in the integrated environment in lieu of directly integrating the
more computationally expensive propulsion system tool directly. Coefficient sets were
obtained by querying multiple engines that had been sized for different shaft-power
implemented and the effects of the engine association logic on the performance of
296
10.2 Conclusions
A series of experiments was performed in order to test and verify certain capabilities
of the integrated subsystem sizing and analysis environment. These included the
following (for each of three chosen baselines):
(Expt. 1.2)
analysis tool, but without directly integrating the tool (Expt. 2.2)
297
8. Ability to assess the relative performance of a large number of subsystem
architectures within an architectural design space relative to a conventional
As part of Expt. 1.1, the performance of an architecture with electric ECS was
Generation and Distribution System (HPGDS) and the Pneumatic Power Generation
and Distribution System (PPGDS), and in certain cases the elimination of these
subsystems altogether. However, significant mass was added back to the aircraft
3. the fact that electric actuation solutions were in general heavier than
the hydraulic solutions that they replaced with the assumed technological
state-of-the-art
In most cases, the net consequence of these counter-acting effects was an increase in
the Operating Empty Weight (OEW) of the MEA architectures. Despite this, the
MEA architectures in many cases showed a reduction in fuel consumption relative to
the conventional architecture. This was largely driven by the elimination of pneumatic
off-takes (bleed air) from the engines through electrification of major subsystems such
298
As part of Expt. 2.1, the heuristic architecting algorithm was tested with
inputs corresponding to existing aircraft with conventional and MEA subsystem
and power sources. Further, when provided inputs corresponding to two existing
MEA, the Airbus A380 and the Boeing 787, the algorithm was able to capture the
elimination of one hydraulic system from the former and the significant expansion of
As part of Expt. 2.2, a higher-fidelity engine model that had been developed
using the Numerical Propulsion System Simulation (NPSS) tool was used as the
truth model. For multiple engines sized to different combinations of shaft-power and
bleed off-takes, the truth model was queried in order to determine coefficients that
represented the degradation in engine TSFC on account of secondary power off-takes.
Rather than determining the penalty coefficients as functions of altitude and Mach
number, they were instead scheduled with respect to normalized mission time, under
TSFC with thrust for the range of variations in thrust between different subsystem
architectures. Penalty coefficients obtained in this manner were used to evaluate SSA
subsystem architecture performance subsequently as part of Expt 4.0.
As part of Expt. 2.3, the net subsystem architecture impact was decomposed
into contributions from individual subsystems and further decomposed into the
contributions arising from mass, secondary power, and drag increments. From the first
level of decomposition, it was found that the ECS (whether pneumatic or electric) had
magnitude. From the second level of decomposition, it was found that the Total
299
Fuel Impact (TFI) of the actuation functions originated almost completely from the
masses of these subsystems. For pneumatic ECS, bleed air off-takes accounted for
the majority of the subsystem’s TFI, whereas for electric ECS, the TFI was more
evenly contributed to by the subsystem mass, shaft-power off-takes, and ram drag.
System (ETS) revealed that such a system would only be feasible for the SSA (flying
short-haul flights) but impractical for the LTA and the VLA (flying long-haul flights).
As part of Expt. 3.1, when strategically chosen model parameters were varied
ECS and IPS (i.e., bleedless architectures) were seen to maintain a fuel consumption
advantage. On the other hand, for architectures that featured All Electric actuation
epistemic uncertainty.
As part of Expt. 3.2, when an advancement of the technological state-of-the-art of
electric motors, power electronics, and electric generator technology was simulated,
all MEA architectures showed improvement in performance. These gains were much
more significant for architectures where these components were more pervasive or
where their rated capacities were higher. MEA architectures featuring higher levels
of electrification were also seen to benefit if the available capacity of the Auxiliary
Power Unit (APU) generators was factored into the sizing of the capacities of the
Engine Driven Generators (EDGs). However, in reality, the feasibility of this would be
As part of Expt 4.0, the performances of 511 MEA architectures were compared
relative to a conventional subsystem architecture for the SSA, LTA, and VLA
with similar observations from several previous research studies and programs. An
300
exception to this trend was observed for the Electric Taxiing System (ETS), which
was established to be practical only for the SSA but not for the LTA or the VLA.
In case of the LTA and the VLA, the cross-over range, beyond which the in-flight
fuel penalty due to the ETS mass exceeded the fuel saved during taxiing, was seen
to be less than the trip distances that aircraft of such sizes are typically operated
on. When the architectures were evaluated using a constant coefficient penalty for
secondary power off-takes, they were found to appear in four clusters characterized
by whether the ECS was electrified and whether the ETS was installed. However,
when the SSA architectures were evaluated using the scheduled penalty coefficients
found to be somewhat more complex. An explanation for this observed behavior was
aircraft sizes with a level of resolution deemed acceptable for the early design phases,
and thus partially or completely supported all hypotheses associated with the three
In order to size and evaluate the performance of the different subsystem solutions
considered within the scope of this dissertation, a number of models were developed
that were computationally inexpensive and required only information that is either
available or easily estimable during the early design phases. Further refinement and
validation of these models will serve to reduce the effect of epistemic uncertainty on
301
to the above, some specific avenues for further research are identified as follows:
(pneumatic) and electric ECS solutions. This is not necessarily the optimal
choice for an electric ECS solution. Given the observed impact of ECS
power electronics equipment becomes significant and increases the ECS heat
load. While the effect of this was implicitly evaluated through a sensitivity
study, the development of a more detailed thermal model of the vehicle and a
requirements into the determination of the engine cycle parameters, which was
The above areas form potential avenues for future research and further extension
of the capabilities of the integrated subsystem sizing and analysis tool. However,
the integrated environment in its current form was sufficient to address the identified
302
APPENDIX A
303
Figure 63: Summary of effects of conventional to All Electric subsystem architecture
transition for LTA (LTA-73310 relative to LTA-00000)
304
APPENDIX B
∆ Block Fuel vs. ∆ Ramp Weight ∆ Block Fuel vs. ∆ Op. Empty Weight
1.5 LTA−00000
2 LTA−00000
1 LTA−73310
LTA−03310
1
0.5
0 0
−0.5
−1 LTA−73310
−1
LTA−70000 −2 LTA−70000
−1.5
LTA−03310
−4 −2 0 2 4 −4 −2 0 2 4
% ∆ Block Fuel % ∆ Block Fuel
∆ Block Fuel vs. ∆ Ramp Weight ∆ Block Fuel vs. ∆ Op. Empty Weight
5
VLA−73310
3 VLA−70000
% ∆ Operating Empty Weight
VLA−70000
% ∆ Max Ramp Weight
2 2.5
VLA−73310
0 0
VLA−00000
VLA−03310 VLA−00000
−1
−2 VLA−03310 −2.5
−3
−5
−6 −4 −2 0 2 4 6 −6 −4 −2 0 2 4 6
% ∆ Block Fuel % ∆ Block Fuel
305
Figure 67: Impact of individual epistemic uncertainty K-factors on performance of
LTA-00000, LTA-70000, LTA-03310, and LTA-73310
306
Figure 68: Impact of individual epistemic uncertainty K-factors on performance of
VLA-00000, VLA-70000, VLA-03310, and VLA-73310
307
B.2 Sensitivity to Technological State-of-the-Art
(Expt. 3.2)
∆ Block Fuel vs. ∆ Ramp Weight ∆ Block Fuel vs. ∆ Op. Empty Weight
4
LTA−73311
1 2 LTA−73310
LTA−73311
LTA−03310
LTA−70000 LTA−70000
0 0
LTA−03310
−1 −2
LTA−73310
−2
−4
−4 −2 0 2 4 −4 −2 0 2 4
% ∆ Block Fuel % ∆ Block Fuel
∆ Block Fuel vs. ∆ Ramp Weight ∆ Block Fuel vs. ∆ Op. Empty Weight
3 VLA−73311
5
% ∆ Operating Empty Weight
2
% ∆ Max Ramp Weight
VLA−73311 VLA−73310
1
VLA−70000 VLA−03310 VLA−70000
0 0
−1 VLA−73310
VLA−03310
−2
−5
−3
−5 0 5 −5 0 5
% ∆ Block Fuel % ∆ Block Fuel
308
Figure 71: Impact of individual technological SOTA K-factors on performance of
LTA-70000, LTA-03310, LTA-73310, and LTA-73311
309
Figure 72: Impact of individual technological SOTA K-factors on performance of
VLA-70000, VLA-03310, VLA-73310, and VLA-73311
310
REFERENCES
[1] Faleiro, L., “Beyond the More Electric Aircraft,” AIAA Aerospace America,
September, 2005 , September 2005, pp. 35–40.
[3] Jones, R., “The More Electric Aircraft - Assessing the Benefits,” Proceedings
of the Institution of Mechanical Engineers, Part G, Journal of Aerospace
Engineering, Vol. 216, 2002, pp. 259–270.
[5] Woodford, C., “Electrics for Aircraft,” Journal of the Royal Aeronautical
Society, Vol. 49, 1945, pp. 518–565.
[6] Jones, R., “The More Electric Aircraft: the past and the future?” IEE
Colloquium on Electrical Machines and Systems for the More Electric Aircraft,
1999, pp. 1/1–1/4.
[7] Rosero, J., Ortega, J., Aldabas, E., and Romeral, L., “Moving Towards a More
Electric Aircraft,” Aerospace and Electronic Systems Magazine, IEEE , Vol. 22,
No. 3, March 2007, pp. 3–9.
[8] Cronin, M., “The all-electric aircraft,” IEE Review , Vol. 36, No. 8, sep 1990,
pp. 309 –311.
[9] Cloyd, J., “Status of the United States Air Force’s More Electric Aircraft
initiative,” Aerospace and Electronic Systems Magazine, IEEE , Vol. 13, No. 4,
April 1998, pp. 17–22.
[11] Rubertus, D., Hunter, L., and Cecere, G., “Electromechanical Actuation
Technology for the All-Electric Aircraft,” IEEE Transactions on Aerospace and
Electronic Systems, Vol. AES-20, No. 3, 1984, pp. 243–249.
[12] Treacy, J., “Flight Safety Issues of an All-Electric Aircraft,” IEEE Transactions
on Aerospace and Electronic Systems, Vol. AES-20, No. 3, 1984, pp. 227–233.
311
[13] Hunt, E., Reid, D., Space, D., and Tilton, F., “Commercial Airliner
Environmental Control System - Engineering Aspects of Air Quality,” Aerospace
Medical Association Annual Meeting, Anaheim, California, 1995.
[14] Sinnett, M., “Boeing 787 No-Bleed Systems: Saving Fuel and Enhancing
Operational Efficiencies,” Boeing Aero Magazine, Quarter 4, Online: http://
www.boeing.com/commercial/aeromagazine/, 2007.
[17] Biedermann, O. and Geerling, G., “Power Control Units with Secondary
Control Hydraulic Motor - A New Concept for Application in Aircraft High
Lift Systems,” Proceedings of the Conference on Recent Advances in Aerospace
Hydraulics, Toulouse, France, 1998.
[18] Churn, P., Maxwell, C., Schofield, N., Howe, D., and Powell, D.,
“Electro-hydraulic actuation of primary flight control surfaces,” IEE
Colloquium on All Electric Aircraft, 1998.
[19] Thompson, K., “Notes on ’The electric control of large aeroplanes’,” Aerospace
and Electronic Systems Magazine, IEEE , Vol. 3, No. 12, December 1988, pp. 19
–24.
[21] Jensen, S., Jenney, G., and Dawson, D., “Flight test experience with an
electromechanical actuator on the F-18 Systems Research Aircraft,” 19th Digital
Avionics Systems Conference, 2000 , Vol. 1, 2000, pp. 2E3/1 – 2E310.
[22] Burkhard, A. and Dietrich, R., “Joint Strike Fighter Integrated Subsystems
Technology, A Demonstration for Industry, by Industry,” Journal of Aircraft,
Vol. 40, 2003, pp. 906–913.
[23] Vohnout, S., Bodden, D., Kim, B., Wagoner, R., Kunst, N., Edwards, P.,
Gleeson, B., Cascio, D., Brzuszkiewicz, S., Wagemans, R., Rounds, M., and
Clements, N., “Prognostic-Enabling of an Electrohydrostatic Actuator (EHA)
System,” Annual Conference of Prognostics and Health Management Society,
Minneapolis, MN, 2012.
312
[24] Van den Bossche, D., “The A380 Flight Control Electrohydrostatic Actuators,
Achievements and Lessons Learnt,” 25th International Congress of the
Aeronautical Sciences, ICAS 2006 , Hamburg, Germany, 2006.
[25] Kulshreshtha, A. and Charrier, J., “Electric Actuation for Flight and Engine
Control: Evolution and Challenges,” SAE-ACGSC Meeting 99 , SAE, Feb 28 -
Mar 2, 2007.
[26] Atallah, K., Caparrelli, F., Bingham, C., Schofield, N., Howe, D., Mellor, P.,
Maxwell, C., Moorhouse, D., and Whitley, C., “Permanent magnet brushless
drives for aircraft flight control surface actuation,” Electrical Machines and
Systems for the More Electric Aircraft (Ref. No. 1999/180), IEE Colloquium
on, London, U.K., 1999, pp. 8/1 –8/5.
[27] Fronista, G. and Bradbury, G., “An electromechanical actuator for a transport
aircraft spoiler surface,” Energy Conversion Engineering Conference, 1997.
IECEC-97., Proceedings of the 32nd Intersociety, Vol. 1, 1997, pp. 694–698.
[28] Recksiek, M., “Advanced High Lift System Architecture with Distributed
Electrical Flap Actuation,” Aviation System Technology Workshop, Hamburg,
Germany, 2009.
[29] Bennett, J., Mecrow, B., Jack, A., Atkinson, D., Sheldon, S., Cooper, B.,
Mason, G., Sewell, C., and Cudley, D., “A prototype electrical actuator for
aircraft flaps and slats,” Electric Machines and Drives, 2005 IEEE International
Conference on, San Antonio, TX, 2005, pp. 41–47.
[31] Bennett, J., Mecrow, B., Atkinson, D., and Atkinson, G., “Safety-critical design
of electromechanical actuation systems in commercial aircraft,” Electric Power
Applications, IET , Vol. 5, No. 1, 2011, pp. 37–47.
[32] Nicolas, Y., “eTaxi - Taxiing aircraft with engines stopped,” Flight
Airworthiness Support Technology (FAST) 51, Airbus Technical Magazine,
http://www.airbus.com/support/publications/, January 2013.
[33] Secunde, R., Macosko, R., and Repas, D., “Integrated Engine-Generator
Concept for Aircraft Electric Secondary Power,” Technical Memorandum
NASA-TM-X-2579, NASA Lewis Research Center, Cleveland, OH, USA, June
1972.
[34] Cao, W., Mecrow, B., G.J., A., Bennet, J., and Atkinson, D., “Overview of
Electric Motor Technologies Used for More Electric Aircraft (MEA),” IEEE
Transactions on Industrial Electronics, Vol. 59, September 2012, pp. 3523–3531.
313
[35] Atkinson, G., Mecrow, B., Jack, D., Atkinson, D., Sangha, P., and Benarous,
M., “The analysis of losses in high-power fault-tolerant machines for aerospace
applications,” IEEE Trans. Ind. Appl., Vol. 42, No. 5, 2006, pp. 1162–1170.
[36] Nam, T., A Generalized Sizing Method for Revolutionary Concepts under
Probabilistic Design Constraints, Ph.D. thesis, School of Aerospace Engineering,
Georgia Institute of Technology, 2007.
[37] Raymer, D., Aircraft Design: A Conceptual Approach, AIAA Education Series,
4th ed., 2006.
[38] Roskam, J., Airplane Design Part V - Component Weight Estimation, Design
Analysis & Research, 1999.
[40] Heimbold, R., Cronin, M., and Howison, W., “Application of Advanced
Electric/Electronic Technology to Conventional Aircraft,” NASA Contractor
Report NASA-CR-163576, Lockheed-California Co., Burbank, CA,
Lockheed-Georgia Co., AiResearch Manufacturing Co., Honeywell Inc.,
July 1980.
[42] Tagge, G., Irish, L., and Bailey, A. R., “Systems Study for an Integrated
Digital/Electric Aircraft (IDEA).” NASA Contractor Report 3840, January
1985.
[43] Cronin, M., Hays, A., Green, F., Radovcich, N., Helsley, C., and Rutchik,
W., “Integrated Digital/Electric Aircraft Concepts Study,” NASA Contractor
Report 3841, 1985.
[44] Hoffman, A., Hansen, I., Beach, R., Plencher, R., Dengler, R., Jefferies, K.,
and Frye, R., “Advanced Secondary Power System for Transport Aircraft,”
NASA Technical Paper 2463, National Aeronautics and Space Administration
(NASA), 1985.
[45] Crowder, R. and Jones, R., “Collaborative Research Initiative into Secondary
Power Systems - Final Summary Report,” CoA CIT and University of
Southampton, October 1992.
[46] Jones, R., “Collaborative Research Initiative into Secondary Power Systems -
Executive Summary Final Report,” CoA CIT, January 1993.
314
[47] TEOS Forum Organization Committee, “Technology for Energy Optimized
Aircraft Equipment Systems - Technical Summary,” Power Optimized Aircraft
(POA) project consortium, 2006.
[48] Dodds, G. and Faleiro, L., “Technology for Energy Optimized Aircraft
Equipment Systems - The Way Forward,” Power Optimized Aircraft (POA)
project consortium, 2006.
[50] Bodie, M., Russell, G., McCarthy, K., Lucuc, E., Zumberge, J., and Wolff,
M., “Thermal Analysis of an Integrated Aircraft Model,” 48th AIAA Aerospace
Sciences Meeting Including the New Horizons Forum and Aerospace Exposition,
2010.
[51] Walters, E.A. and Iden, S. et al., “INVENT Modeling, Simulation, Analysis
and Optimization,” 48th AIAA Aerospace Sciences Meeting Including the New
Horizons Forum and Aerospace Exposition, 2010.
[52] Jackson, D., Robust Aircraft Subsystem Conceptual Architecting, Ph.D. thesis,
Georgia Institute of Technology, 2013.
[53] de Tenorio, C., Method for Collaborative Conceptual Design of Aircraft Power
Architectures, Ph.D. thesis, Georgia Institute of Technology, Atlanta, GA, USA,
2010.
[55] Armstrong, M., A Process for Function Based Architecture Definition and
Modeling, Master’s thesis, Georgia Institute of Technology, 2008.
[57] Liscouet-Hanke, S., Mare, J.-C., and Pufe, S., “Simulation Framework for
Aircraft Power System Architecting,” Journal of Aircraft, Vol. 46, No. 4, 2009,
pp. 1375–1380.
[58] Liscouet-Hanke, S., Pufe, S., and Mare, J.-C., “A Simulation Framework for
Aircraft Power Management,” Proceedings of the Institution of Mechanical
Engineers, Part G: Journal of Aerospace Engineering, Vol. 222, No. 6, June
2008, pp. 749–756.
315
[59] Seresinhe, R. and Lawson, C., “Electrical load-sizing methodology to aid
conceptual and preliminary design of large commercial aircraft,” Proceedings
of the Institution of Mechanical Engineers, Part G, Journal of Aerospace
Engineering, May 2014, pp. 1–22.
[60] Seresinhe, R., Impact of aircraft systems within aircraft operation: A MEA
trajectory optimization study, Ph.D. thesis, Centre for Aeronautics, School
of Aerospace, Transport, and Manufacturing, Cranfield University, September
2014.
[61] Leonard, J., “The All-Electric Fighter Airplane Flight Control Issues,
Capabilities, and Projections,” IEEE Transactions on Aerospace and Electronic
Systems, Vol. AES-20, 3, May 1984, pp. 234–242.
[62] Hlavacek, I., Chleboun, J., and Babuska, I., Uncertain Input Data Problems
and the Worst Scenario Method , North-Holland Series in Applied Mathematics
and Mechanics, Elsevier B.V., 2004.
[63] Moir, I. and Seabridge, A., Design and Development of Aircraft Systems, No.
ISBN: 978-1-119-94119-4, Wiley Aerospace Series, 2013.
[65] McCullers, L., Flight Optimization System, Release 8.11, User’s Guide, NASA
Langley Research Center, Hampton, VA 23681-0001, October 9 2009.
[66] Rudolph, P., “High Lift Systems on Subsonic Commercial Airliners,” NASA
Contractor Report 4746 CONTRACT A46374D(LAS), National Aeronautics
and Space Administration, September 1996 1996.
[67] Bennett, J., Fault Tolerant Electromechanical Actuators for Aircraft, Ph.D.
thesis, Newcastle University, School of Electrical, Electronic and Computer
Engineering, 2010.
[68] Bennett, J., Mecrow, B., Atkinson, D., Maxwell, C., and Benarous, M.,
“Fault-tolerant electric drive for an aircraft nose wheel steering actuator,” IET
Electrical Systems in Transportation, Vol. 1, No. 3, 2011, pp. 117–125.
[69] Bennett, J., Mecrow, B., Jack, A., and Atkinson, D., “A Prototype Electrical
Actuator for Aircraft Flaps,” Industry Applications, IEEE Transactions on,
Vol. 46, No. 3, 2010, pp. 915–921.
[70] Botten, S., Whitley, C., and King, A., “Flight Control Actuation Technology for
Next-Generation All-Electric Aircraft,” Technology Review Journal - Millenium
Issue, 2000, pp. 55–68.
316
[71] Heinrich, A., Ross, R., Zumwalt, G., Provorse, J., Padmanabhan, V.,
Thompson, J., and Riley, J., “Aircraft Icing Handbook - Volume 2 of
3,” DOT/FAA/CT-88/8- I, Department of Transporation, Federal Aviation
Administration Technical Center, 1991.
[72] Scholz, D., “Fuel Costs due to Aircraft Systems - Calculated from Small Time
Intervals,” Memo, Aircraft Design and Systems Group, Hamburg University of
Applied Sciences, 2007.
[73] Blanding, D., “Subsystem Design and Integration for the More Electric
Aircraft,” 25th International Congress of the Aeronautical Sciences (ICAS
2006), Hamburg, Germany, September 3-8, 2006.
[74] Nelson, T., “787 Systems and Performance,” Flight Operations Engineering,
Boeing Commercial Airplanes, March 2009, online: http://myhres.com/
Boeing-787-Systems-and-Performance.pdf, accessed November 4, 2015.
[76] Chakraborty, I., Trawick, D., Hegde, C., Choi, H., Mendez-Ramos, E., and
Mavris, D. N., “Development of a Modeling and Simulation Environment for
Real-time Performance Analysis of Electric Actuators in Maneuvering Flight,”
51st AIAA Aerospace Sciences Meeting Including The New Horizons Forum and
Aerospace Exposition, No. AIAA-2013-0471, Grapevine, TX, January 2013.
[78] Scholz, D., “Development of a CAE-Tool for the Design of Flight Control and
Hydraulic Systems,” AeroTech ’95 , Birmingham, U.K., 1995.
[79] Scholz, D., “Equations for a Preliminary Actuator Design,” Technical Note
TN-EV52-360/91, Deutsche Airbus, 1991.
[80] Scholz, D., “Computer Aided Engineering for the Design of Flight Control and
Hydraulic Systems,” ICAS Proccedings 1996 (20th Congress of the International
Council of the Aeronautical Sciences), 1996.
[81] Airbus Training, “Airbus A330 Flight Crew Operating Manual - Flight
Controls,” online: http://www.smartcockpit.com/aircraft-ressources/
A330-Flight_Controls.html, accessed November 4, 2015.
[82] Boeing 777 Aircraft Maintenance Manual, Chapter 27 - Flight Controls, 2006.
317
[84] Society of Automotive Engineers, “SAE Aerospace Information Report (AIR)
5005A - Commercial Aircraft Hydraulic Systems,” 2010.
[85] Socheleau, J., Mare, J.-C., and Baudu, P., “Actuation Technologies and
Application - Flight Controls and Thrust Reverser Actuation,” Technologies for
Energy Optimized Aircraft Equipment Systems (TEOS forum), Paris, France,
2006.
[86] Chakraborty, I., Mavris, D., Emeneth, M., and Schneegans, A., “A
Methodology for Vehicle and Mission Level Comparison of More Electric
Aircraft Subsystem Solutions - Application to the Flight Control Actuation
System,” Proceedings of the Institution of Mechanical Engineers, Part G:
Journal of Aerospace Engineering, Vol. 229, No. 6, July 28 2014, pp. 1088–1102.
[87] Chakraborty, I., Mavris, D., Emeneth, M., and Schneegans, A., “An Integrated
Approach to Vehicle and Subsystem Sizing and Analysis for Novel Subsystem
Architectures,” Proceedings of the Institution of Mechanical Engineers, Part G:
Journal of Aerospace Engineering, DOI: 10.1177/0954410015594399, accepted
for publication July 20, 2015.
[88] Garmendia, D., Chakraborty, I., and Mavris, D., “A Method for Evaluating
Electrically Actuated Redundant Control Surface Layouts for the Hybrid Wing
Body,” AIAA Journal of Aircraft, DOI: 10.2514/1.C033061, accepted for
publication Feb 18, 2015.
[89] Garmendia, D., Chakraborty, I., and Mavris, D., “A Multidisciplinary Approach
to Assessing Actuation Power of a Hybrid Wing Body,” AIAA Journal of
Aircraft, accepted for publication Sep 8, 2015.
[90] Chakraborty, I., Trawick, D., Jackson, D., and Mavris, D., “Electric Control
surface Actuator Design Optimization and Allocation for the More Electric
Aircraft,” AIAA Aviation 2013 Conference, No. AIAA 2013-4283, Los Angeles,
California, August 12-14, 2013.
[91] Rea, J., “Boeing 777 High Lift Control System,” IEEE Aerospace and Electronic
Systems Magazine, Vol. 8, No. 8, August 1993, pp. 15–21.
[92] Eaton Aerospace Group, “Aerospace Capabilities - Boeing 777,” online:
http://www.eaton.com/ecm/groups/public/@pub/@eaton/@aero/
documents/content/ct_194202.pdf, accessed November 4, 2015.
[93] Martinez, M., Sawata, T., Rouge-Carrassat, T., and Blineau, J.-M., “Electrical
Power Sources and Aircraft Power Networks,” Technologies for Energy
Optimized Aircraft Equipment Systems (TEOS forum), Paris, France, 2006.
[94] Slingerland, R., Zandstra, S., Scholz, D., and Seeckt, K., “Green
Freighter Systems,” 46th AIAA Aerospace Sciences Meeting and Exhibit, No.
AIAA-2008-146, American Institute of Aeronautics and Astronautics, 7-10
January 2008.
318
[95] Li, W. and Fielding, J., “Preliminary Study of EMA Landing Gear Actuation,”
28th International Congress of the Aeronautical Sciences (ICAS 2012),
Brisbane, Australia, 2012.
[96] Jenkins, S., “Landing Gear Design and Development,” Proceedings of the
Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering,
Vol. 203, No. 1, 1989, pp. 67–73.
[97] Currey, N., Aircraft Landing Gear Design: Principles and Practices, No. ISBN:
978-0-930403-41-6, AIAA Education Series, American Institute of Aeronautics
and Astronautics, 1988.
[98] Young, D., “Aircraft landing gears - the past, present and future,” Proceedings
of the Institution of Mechanical Engineers, Part D: Journal of Automobile
Engineering, Vol. 200, No. 2, 1986, pp. 75–92.
[99] Roskam, J., Airplane Design Part IV - Layout of Landing Gear and Systems,
Design Analysis & Research, 1999.
[100] Grac, S., “Distributed and Redundant Electro-mechanical Nose Wheel Steering
System (DRESS) - Publishable Executive Summary,” Messier-Bugatti (Safran
Group), 2010.
[102] Hendricks, E. and Tong, M., “Performance and Weight Estimates for an
Advanced Open Rotor Engine,” NASA/TM2012-217710, National Aeronautics
and Space Administration, Glenn Research Center, Cleveland, OH, September
2012.
[103] Nevoret, P., “Recent Advances and Future Electrical Landing Gear Systems,”
ICAS Biennial Workshop - ”The More Electrical Aircraft: Achievements &
perspectives for the future”, Cape Town, South Africa, September 2013.
[104] Wahi, M., Warren, S., and Straub, H., “An Extended Prediction Model for
Airplane Braking Distance and a Specification for a Total Braking Prediction
System,” Technical Report ASD-TR-77-6, Boeing Commercial Airplane
Company, Aeronautical Systems Division, Air Force Systems Command,
Wright-Patterson Air Force Base, OH 45433, 1977.
[105] The Goodyear Tire and Rubber Company, “Aircraft Tire Data Book,” 1144
East Market Street, Akron, OH 44316-0001, 2002.
[106] Collins, A., “EABSYS: Electrically Actuated Braking System,” IEE Colloquium
on Electrical Machines and Systems for the More Electric Aircraft, London,
U.K., 1999, p. 4.
319
[107] Root, R., “Brake Energy Considerations in Flight Operations,” Flight
Operations Engineering, Boeing Commercial Airplanes, September, 2003,
online: www.smartcockpit.com, accessed: November 4, 2015.
[108] Allen, T., Miller, T., and Preston, E., “Operational Advantages of Carbon
Brakes,” Boeing Aero Magazine, Quarter 3, Online: http://www.boeing.com/
commercial/aeromagazine/, 2009.
[110] Chai, S. and Mason, W., “Landing Gear Integration in Aircraft Conceptual
Design,” 6th AIAA/NASA/ISSMO Symposium on Multidisciplinary Analysis
and Optimization, No. AIAA 96-4038, Bellevue, WA, September 4-6 1996.
[114] Chakraborty, I., LeVine, M., and Hassan, M., “Assessing Taxiing Trade
Spaces from Aircraft, Airport, and Airline Perspectives,” AIAA AVIATION
2015 Conference, No. AIAA-2015-2386, American Institute of Aeronautics and
Astronautics, Dallas, TX, June 22-26 2015.
[115] Morris, K., “Results from a number of surveys of power settings used during taxi
operations,” Technical Report EJT/KMM/1126/14.8, British Airways, October
2005.
[116] Sinnett, M., “787 Program - Electrical System and Batteries,” Online:
www.boeing.com/787-media-resource/docs/Sinnett-TOS-Deck.pdf, February
2013.
[118] Colin, A., Herzog, J., Dodds, G., and Larue, F., “Environmental Control System
and Wing Ice Protection (ECS & WIPS),” Technologies for Energy Optimized
Aircraft Equipment Systems (TEOS forum), Paris, France, 2006.
320
[119] Muller, C., Scholz, D., and Giese, T., “Dynamic Simulation of Innovative
Aircraft Air Conditioning,” First CEAS European Air and Space Conference,
Berlin, September 10-13 2007.
[120] Pratt, J., Klebanoff, L., Munoz-Ramos, K., Akhil, A., Curgus, D., and
Schenkman, B., “Proton Exchange Membrane Fuel Cells for Electrical Power
Generation On-Board Commercial Airplanes,” Sandia Report SAND2011-3119,
Sandia National Laboratories, Albuquerque, New Mexico 87185 and Livermore,
California 94550, May 2011.
[121] Airbus Industrie, “Airbus A320 Simulator Flight Crew Operating Manual
- Air Conditioning, Pressurization, and Ventilation,” online: http://www.
smartcockpit.com/aircraft-ressources/A320-Air_Conditioning_and_
Pressurization.html, accessed November 4, 2015.
[122] Airbus Industrie, “A319/A320/A321 Technical Training Manual, 21 Air
Conditioning,” 2000.
[123] Airbus Industrie, “Airbus A330 Simulator Flight Crew Operating Manual
- Air Conditioning, Pressurization, and Ventilation,” online: http://www.
smartcockpit.com/aircraft-ressources/A330-Aircond-Press-Vent.
html, accessed November 4, 2015.
[124] Airbus Industrie, “Airbus A340 Simulator Flight Crew Operating
Manual - Air Conditioning, Pressurization, and Ventilation,” online:
http://www.smartcockpit.com/aircraft-ressources/FCOM_A340-Air_
Cond-Press-Vent.html, accessed November 4, 2015.
[125] Boeing, “Boeing 737NG - Systems Summary, Air Systems,” online: http://
www.smartcockpit.com/aircraft-ressources/B_NG-Air_Systems.html,
accessed November 4, 2015.
[126] Brasseur, A., Leppert, W., and Pradille, A., “Inside the 747-8 New
Environmental Control System,” Boeing Aero Magazine, Issue 45, Quarter
1, 2012. Online: http://www.boeing.com/commercial/aeromagazine/
articles/2012_q1/pdfs/AERO_2012q1_article4.pdf.
[127] Hodal, P. and Liu, G., “Bleed Air Temperature Regulation System: Modeling,
Control, and Simulation,” Proceedings of the 2005 IEEE Conference on Control
Applications, Toronto, Canada, August 28-31, 2005.
[128] Heinrich, A., Ross, R., Zumwalt, G., Provorse, J., Padmanabhan, V.,
Thompson, J., and Riley, J., “Aircraft Icing Handbook - Volume 1 of
3,” DOT/FAA/CT-88/8- I, Department of Transporation, Federal Aviation
Administration Technical Center, 1991.
[129] Thomas, S. K., Cassoni, R. P., and MacArthur, C. D., “Aircraft anti-icing and
de-icing techniques and modeling,” Journal of Aircraft, Vol. 33, No. 5, 1996,
pp. 841–854.
321
[130] Bu, X., Lin, G., Yu, J., Shen, X., and Hou, P., “Numerical analysis of a swept
wing hot air ice protection system,” Proceedings of the Institution of Mechanical
Engineers, Part G: Journal of Aerospace Engineering, Vol. 227, No. 10, 2013,
pp. 1608–1622.
[131] Al-Khalil, K., Keith, T., DeWitt, K., Nathman, J., and Dietrich, D., “Thermal
Analysis of Engine Inlet Anti-Icing Systems,” Journal of Propulsion, Vol. 6,
No. 5, 1990, pp. 628–634.
[132] Patricelli, L., Innovative Solutions for the Thermal Control of Aeronautic
Vehicles, Ph.D. thesis, Dipartimento di Ingegneria Industriale, Universit degli
Studi di Padova (University of Padua), 2014.
[135] Jeck, R., “Icing Design Envelopes (14 CFR Parts 25 and 29, Appendix C)
Converted to a Distance-Based Format,” Final Report DOT/FAA/AR-00/30,
Federal Aviation Administration, Airport and Aircraft Safety Research and
Development, William J. Hughes Technical Center, 2002.
[137] Bowden, D., Gensemer, A., and Skeen, C., “Engineering Summary of Aircraft
Icing Technical Data,” Technical Report ADS-4, Federal Aviation Agency,
Washington D.C., 1964.
[138] Wright, W., “An Evaluation of Jet Impingement Heat Transfer Correlations for
Piccolo Tube Application,” 42nd Aerospace Sciences Meeting and Exhibit, No.
AIAA-2004-0062, 2004.
[139] Al-Khalil, K., Horvath, C., Miller, D., and Wright, W., “Validation of Thermal
Ice Protection Computer Codes: Part 3 - The Validation of ANTICE,” 35th
Aerospace Sciences Meeting and Exhibit, No. AIAA-97-0051, Reno, Nevada,
1997.
[140] Gray, V., “Correlations Among Ice Measurements, Impingement Rates, Icing
Conditions, and Drag Coefficients for Unswept NACA 65A004 Airfoil,”
NACA Technical Note 4151, National Advisory Committee for Aeronautics,
Washington, February 1958.
322
[141] Sadeghi, T. and Lyons, A., “Fault Tolerant EHA Architectures,” IEEE AES
Systems Magazine, Vol. 7, No. 3, March 1992, pp. 32–42.
[142] Wheeler, P. W., Clare, J. C., Trentin, A., and Bozhko, S., “An overview of the
more electrical aircraft,” Proceedings of the Institution of Mechanical Engineers,
Part G: Journal of Aerospace Engineering, Vol. 227, No. 4, 2012, pp. 578–585.
[143] Giraud, X., Piquet, H., Budinger, M., Roboam, X., Sartor, M., and Vial, S.,
“Knowledge-based system for aircraft electrical power system reconfiguration,”
Electrical Systems for Aircraft, Railway and Ship Propulsion (ESARS), 2012 ,
Oct 2012, pp. 1–6.
[144] Banel-Caule, I., “Systems Weight Estimation Enhanced Method for Early
Project Phases,” 66th Annual Conference of Society of Allied Weight Engineers,
Inc., No. SAWE Paper No. 3430, Society of Allied Weight Engineers, Inc.,
Madrid, Spain, May 26-31 2007.
[145] Airbus Industrie, “Airbus A320 Aircraft Maintenance Manual, ATA29,” 2006.
[146] Xia, X., Dynamic Power Distribution Management for All Electric Aircraft,
Master’s thesis, School of Engineering, Cranfield University, 2011.
[147] Whyatt, G. and Chick, L., “Electrical Generation for More-Electric Aircraft
using Solid Oxide Fuel Cells,” Report PNNL-21382, Pacific Northwest National
Laboratory, Richland, WA 99352, April 2012.
[148] Christou, I., Nelms, A., Cotton, I., and Husband, M., “Choice of Optimal
Voltage for More Electric Aircraft Wiring Systems,” IET Electrical Systems in
Transportation, Vol. 1, No. 1, 2010, pp. 24–30.
[153] Peacock, D., “Effective Design using Titanium,” Intl. Conf. Titanium in
Practical Applications, Norwegian Association of Corrosion Engineers, 1990.
323
[154] Australian Transport Safety Bureau, “Air/pressurization involving a Boeing
Company 767-277. VH-RMF, near Canberra Aerodrome, ACT on 21 October
1993,” Investigation number: 199303375, 1993, online: https://www.atsb.
gov.au/publications/investigation_reports/1993/aair/199303375.
aspx, accessed November 4, 2015.
[156] European Aviation Safety Agency (EASA), EASA Type Certification Data
Sheet - CFM International S.A. - CFM56-7B series engines, December 17 2012.
[157] Scholz, D., Seresinhe, R., Staack, I., and Lawson, C., “Fuel Consumption due
to Shaft Power Off-takes from the Engine,” 4th International Workshop on
Aircraft System Technologies, AST 2013 , Shaker, Hamburg,, April 23-24 2013,
pp. 169–179.
[158] Kirby, M. and Mavris, D., “The Environmental Design Space,” 26th
International Congress of the Aeronautical Sciences (ICAS), No.
ICAS-2008-4.7.3, Anchorage, Alaska, September 14-19 2008.
[159] Schutte, J., Tai, J., and Mavris, D., “Multi-Design Point Cycle
Design Incorporation into the Environmental Design Space,” 48th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, Atlanta,
GA, July 30 - Aug 1 2012.
[160] Schutte, J., Tai, J., Sands, J., and Mavris, D., “Cycle Design Exploration
using Multi-Design Point Approach,” Proceedings of ASME Turbo Expo 2012 ,
Copenhagen, Denmark, June 11-15 2012.
[161] Simsic, C., “Electric actuation system duty cycles,” Aerospace and Electronics
Conference, 1991. NAECON 1991., Proceedings of the IEEE 1991 National ,
1991, pp. 540–545 vol.2.
[162] Blanding, D., Sexton, M., Segal, M., and Grow, J., “The Application of
Confidence Interval in the Evaluation of Electric Actuation Duty Cycle,” 27th
Congress of International Council of the Aeronautical Sciences (ICAS), No.
ICAS 2010-7.3.3, Nice, France, 19 - 24 September 2010.
[163] Society of Automotive Engineers, “Aircraft Fuel Weight Penalty Due to Air
Conditioning, SAE Aerospace Applied Thermodynamics Manual - SAE AIR
1168/8,” 1989.
324
VITA
Imon Chakraborty hails from Kolkata, West Bengal, India. He enrolled in the
undergraduate program at the National Institute of Technology, Tiruchirappalli
his Doctor of Philosophy (Ph.D.) degree in December 2015. During the course
Single-Engine Land) certification, and at the time of writing is checked out on the
Cessna 172 and Diamond DA20 aircraft. He is the recipient of the 2015 William
T. Piper, Sr. General Aviation Systems Graduate Award. He is a member of the
American Institute of Aeronautics and Astronautics (AIAA) and the Aircraft Owners
325